WO2005084036A1 - 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム - Google Patents
動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム Download PDFInfo
- Publication number
- WO2005084036A1 WO2005084036A1 PCT/JP2005/002786 JP2005002786W WO2005084036A1 WO 2005084036 A1 WO2005084036 A1 WO 2005084036A1 JP 2005002786 W JP2005002786 W JP 2005002786W WO 2005084036 A1 WO2005084036 A1 WO 2005084036A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pixel
- motion vector
- correlation
- representative point
- unit
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/537—Motion estimation other than block-based
- H04N19/54—Motion estimation other than block-based using feature points or meshes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/223—Analysis of motion using block-matching
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/43—Hardware specially adapted for motion estimation or compensation
Definitions
- Motion vector detection device motion vector detection method, and computer program
- the present invention relates to a motion vector detection device, a motion vector detection method, and a computer program. More specifically, the present invention relates to a motion vector detection device that executes a motion vector detection process from moving image data, a motion vector detection method, and a computer program. Background art
- Motion-compensated image coding in high-efficiency coding of moving image data moving object detection processing in a traffic monitoring system or a visual sensor of an autonomous vehicle, speed detection processing, and the like are included in image data. Processing for detecting the direction and magnitude (velocity) of the motion of each object, that is, processing for detecting a motion vector, is required.
- the MPEG (Moving Picture Coding Experts Group) system which is an international standard system for high-efficiency moving image coding
- DCT Discrete Cosine Transform
- motion-compensated predictive coding the correlation between the current frame constituting moving image data and the image signal level in the continuous frame of the immediately preceding frame is detected, and the motion vector is calculated based on the detected correlation.
- Efficient coding is achieved by performing motion image correction processing based on the obtained and detected motion vectors.
- a block matching method is known. An overview of the block matching method will be described with reference to FIG.
- a temporally continuous frame image constituting a moving image for example, the current frame [F] 20 at time (t) shown in the figure and the current frame [F] 20 before time (t-1)
- One screen of the frame image is a small
- the current frame [F] 20 is set as a reference frame, and a detection block By21 of the reference frame is set to a predetermined
- a detection block with a small value difference that is, a pixel block with the best matching (highest correlation) is detected. If the reference block Bxl l of the previous frame [F] 10 is the current frame [F] 20
- the block matching method is a method of performing a correlation determination (matching determination) between frames in units of a predetermined block (mXn) to obtain a motion vector.
- a motion vector is obtained for each block.
- the correlation value of each block that is, the evaluation value indicating the degree of matching, for example, the value of pixels at the same spatial position between a plurality of pixels in the reference block Bx and a plurality of pixels in the inspection block
- the sum of the absolute values of the frame differences calculated by integrating the absolute values of the calculated frame differences is applied.
- a block when a block includes a moving part and a still part, the movement detected in units of a block does not accurately correspond to the movement of each pixel in the block. ,.
- Such a problem can be adjusted by setting the block size. For example, when the size of the block is increased, the problem of a plurality of motions in the block is likely to occur in addition to the increase in the amount of calculation. Conversely, if the size of the block is reduced so that multiple motions are not included in the block, the matching judgment area becomes smaller, and the accuracy of motion detection decreases. Problems arise. That is, when performing block matching, the possibility that a large number of inspection blocks similar to the reference block, that is, many inspection blocks having a high correlation with the reference block appear.
- Patent Document 1 a motion vector detection method and a detection device capable of detecting a motion vector for each pixel without increasing the amount of calculation and preventing erroneous detection. is suggesting.
- the point of the motion vector detection process disclosed in Patent Document 1 is that the process of the first step, which does not determine the motion vector by calculating the evaluation value for each pixel or block, is performed in one of the frames.
- a plurality of blocks composed of a plurality of pixels are set, a representative point of each block is set, a correlation between each representative point and each pixel of a search area set in the other frame is checked, and an evaluation value based on the correlation information is calculated.
- an evaluation value table as correlation information based on the evaluation value is formed, and a plurality of candidate vectors are extracted from the evaluation value table.
- a candidate vector considered to be the best for each pixel is selected from the extracted candidate vectors and associated with each other, and determined as a motion vector for each pixel. in this way,
- the motion vector of each pixel is obtained by the above processing.
- This method is hereinafter referred to as a candidate vector method.
- the advantage of the motion vector detection process using the candidate vector method is that the amount of calculation can be reduced by extracting a limited number of candidate vectors based on the evaluation value table. In addition, narrow down the boundaries of the subject where motion vectors are likely to be erroneously detected. In some cases, it is possible to determine the best motion vector corresponding to each pixel from among the candidate vectors.
- a motion vector of each pixel is calculated as an evaluation value such as a pixel difference between frames as an evaluation value, and a full search process is performed to obtain an evaluation value for all pixels in the frame. Since the process of determining the best motion vector corresponding to each pixel from the candidate vectors that have been narrowed down in advance can be performed, the probability that the same evaluation value will occur is reduced compared to the full search process, and false detection is performed. Is prevented.
- the representative point of each block is set, the correlation between each representative point and each pixel of the search area set in the other frame is checked, and the correlation information is used. This is a process of calculating an evaluation value based on the evaluation value and integrating the evaluation value.
- the integrated evaluation value is set. That is, if the absolute value of the difference between the representative point pixel X and the input pixel Y included in the search area is equal to or smaller than a certain threshold TH, the integrated evaluation value is set. That is,
- the graph shown on the right side of FIG. 2 shows the pixel level of one line in the X direction passing through the representative point 38 of the previous frame 30 and the pixel level of one line in the X direction in the search area 32 of the current frame. And show.
- the block matching process is a process of setting a pixel near a pixel of interest in a previous frame as a block and detecting an overall correlation of a plurality of pixels included in the block.
- it is necessary to increase the block size and perform accurate correlation determination.
- Increasing the block size increases the amount of calculation of the evaluation value, such as the sum of absolute difference values, which is performed as an operation for calculating the correlation, lowers the efficiency, and increases the memory for holding the pixel values. Must be performed, which causes a problem that the hardware scale becomes large.
- Patent Document 1 JP 2001-61152 A
- the present invention has been made in view of the above-described problems, and is not limited to the generation of the evaluation value table based on the representative point matching processing.
- a motion vector detection device that enables generation of a more accurate evaluation value table by performing a correlation determination process in consideration of a spatial waveform as a motion vector and that can more accurately execute motion vector detection. And a motion vector detection method, and a computer program.
- the evaluation value table in generating the evaluation value table based on the representative point matching process, it is determined whether not only the representative point but also a pixel near the representative point has a motion similar to the representative point, and the correlation is determined.
- a more accurate evaluation value table can be generated, and a motion vector detection device and a motion vector detection method that can more accurately execute a motion vector detection, and a computer '' The purpose is to provide the program.
- the present invention does not apply block matching when determining a motion vector corresponding to each pixel from a plurality of candidate vectors, for example, in a motion vector detection process based on the candidate vector method, and provides an accurate motion vector. It is an object of the present invention to provide a motion vector detection device, a motion vector detection method, and a computer program which enable determination of the motion vector.
- a first aspect of the present invention provides:
- An evaluation value table forming unit that generates an evaluation value table based on pixel value correlation information between different frames on a time axis, which is a motion vector detection device that detects a motion vector from moving image data;
- a motion vector determining unit that detects and associates a motion vector with respect to a frame constituent pixel of the moving image data based on the evaluation value table
- the evaluation value table forming unit
- a pixel correlation calculation unit that calculates correlation information between different frames on the time axis based on a representative point matching process based on a representative point selected from one frame; a calculation result of the pixel correlation calculation unit; The reliability of motion by applying at least one of the calculation results based on the pixel value difference between the pixel and the pixel around the pixel of interest
- a weight calculator for generating an index
- An evaluation value table calculation unit that integrates evaluation values corresponding to the reliability indices calculated by the weight calculation unit and generates an evaluation value table
- a motion vector detecting device A motion vector detecting device.
- the motion vector detecting device further includes a motion similarity detecting unit that detects a motion similarity between the representative point corresponding to the pixel of interest and a representative point near the representative point.
- the weight calculation unit includes, as a detection result of the motion similarity detection unit, a pixel between the representative point and the nearby representative point based on a determination that there is motion similarity between the representative point and the nearby representative point. The feature is to generate a motion reliability index in consideration of the value difference.
- the weight calculation unit includes a condition that a determination that there is motion similarity between representative points is input from the motion similarity detection unit. Based on the brightness level of the focused representative point: Pm, and the brightness level of the N neighboring representative points determined to have similar motion: Pn, the reliability index is given by:
- the motion vector detection device further includes: a determination result of the presence or absence of correlation based on a representative point matching process from the pixel correlation calculation unit; A calculating unit for inputting the reliability index from the calculating unit, wherein the calculating unit adds a reliability determination result based on the correlation index based on the representative point matching process from the pixel correlation calculating unit and the reliability index or It is characterized in that a multiplication process is executed, a final evaluation value is calculated, and the final evaluation value is output to the evaluation value table calculation unit.
- the weight calculation unit comprises:
- the reliability index is set to the following parameter, that is,
- It is characterized in that it is configured to calculate as a value that reflects at least (a) and (c) above, and the deviation force.
- the evaluation value table forming unit further includes a representative point stillness determining unit that determines whether the representative point is in a still area.
- the weight calculation unit is configured to execute a process of setting the value of the reliability index to 0 or decreasing the value when the representative point is determined to be in the stationary region by the representative point stillness determination unit. It is characterized by being.
- the evaluation value table forming unit further generates flag data corresponding to pixel value difference data of the pixel of interest and a pixel in the vicinity of the pixel of interest.
- a flag data calculation unit; and a flag correlation calculation unit that performs a process of calculating flag data correlation information between frames based on the flag data, wherein the weight calculation unit is based on a calculation result of the pixel correlation calculation unit.
- a weighting factor W is calculated by applying at least one of the pixel correlation information and the flag correlation information based on the calculation result of the flag correlation calculation unit, and the reliability as a calculated value based on the weighting factor W is calculated.
- An index is generated, and a new index is generated based on a reliability index taking into account the pixel value difference between the representative point and the neighboring representative point, and a reliability index as a calculated value based on the weight coefficient W.
- the evaluation value table calculating unit generates an evaluation value table by integrating the evaluation values corresponding to the new reliability index K calculated by the weight calculating unit. It is characterized by being.
- flag data corresponding to pixel value difference data between the target pixel and pixels around the target pixel is generated.
- a flag data calculation unit for performing calculation processing of flag data correlation information between frames based on the flag data wherein the weight calculation unit includes a calculation result of the pixel correlation calculation unit.
- the weight coefficient W is calculated by applying at least one of the pixel correlation information based on the weighting factor and the flag correlation information based on the calculation result of the flag correlation calculating unit, and the reliability as a calculated value based on the weighting factor W is calculated. It is characterized by generating a degree index.
- the flag data calculating unit converts the flag data corresponding to the pixel value difference data of the target pixel and the pixel in the vicinity of the target pixel from the pixel value difference data. It is characterized in that it is configured to execute processing for calculating as a small number of bits.
- the weight calculation processing section includes at least a representative point pixel as a weight coefficient calculated from pixel correlation information based on a calculation result of the pixel correlation calculation section. , And a value W calculated based on a difference between the pixel value of the pixel of interest and the pixel value of the correlation determination target pixel including the target pixel and its surrounding pixels is set as a weighting coefficient W.
- the weight calculation processing section includes a pixel correlation information based on a calculation result of the pixel correlation calculation section, and a calculation result of the flag correlation calculation section.
- the flag correlation information power based on the pixel values of the representative point pixel is calculated as the pixel value: X, and the flag data based on the pixel value difference data of two adjacent pixels of the representative point pixel: The value Y, and its two neighboring pixels
- Flag data based on the pixel value difference data of, ⁇ ,,, and the magnitude of the difference between X and ⁇
- the weight calculation processing unit calculates activity A as an index value indicating the complexity of the image data, and calculates the calculated activity A and Processing for calculating a reliability index based on at least a pixel value of a representative point pixel and a weighting coefficient W calculated based on a magnitude of a difference between a pixel value of a correlation determination target pixel including the target pixel and its peripheral pixels. Characterized in that it is configured to execute To do.
- the motion vector detecting device further includes a candidate vector extraction unit that extracts one or more candidate vectors based on the evaluation value table.
- the motion vector determining unit is configured to execute a process of selecting a motion vector corresponding to each frame constituent pixel of moving image data from the candidate vectors and associating the selected motion vector.
- the motion vector determination unit extracts a feature pixel from a region near a pixel of interest to which the motion vector is associated, and performs a correlation determination process based on the feature pixel. Based on this configuration, a motion vector corresponding to the pixel of interest is determined.
- the motion vector determination unit includes a pixel value of a pixel of interest associated with the motion vector, and a pixel value of a pixel specified by the candidate vector. And performing a process of selecting only those having a high correlation from the candidate vectors by performing a correlation determination with the candidate vector.
- the correlation determination process based on the characteristic pixels is performed by the temporary determination unit. It is characterized in that it is configured to be executed as processing targeting only the selected candidate level.
- the motion vector determination unit extracts a feature pixel from a region near the pixel of interest to which the motion vector is associated, and performs a correlation based on the position information of the feature pixel. It is characterized in that the motion vector corresponding to the pixel of interest is determined based on the determination processing.
- the motion vector determination unit extracts a feature pixel from a region near the pixel of interest to which the motion vector is associated, and based on the pixel value information of the feature pixel.
- the feature is that the motion vector corresponding to the pixel of interest is determined based on the correlation determination processing.
- the motion vector determination unit is characterized by a pixel having a maximum absolute value of a pixel value difference between a pixel of interest and a neighboring pixel of a pixel of interest associated with a motion vector.
- the motion vector corresponding to the pixel of interest is determined based on the correlation determination process based on the position information of the characteristic pixel. Is a special feature.
- the motion vector determining section is characterized by a pixel having a maximum pixel value difference absolute value from a pixel in the vicinity of the pixel of interest associated with the motion vector.
- the pixel is extracted as a pixel, and a motion vector corresponding to the pixel of interest is determined based on a correlation determination process based on pixel value information of the characteristic pixel.
- the motion vector determination unit extracts a plurality of feature pixels from a region near a pixel of interest to which the motion vector is associated, and determines the plurality of feature pixels based on the plurality of feature pixels.
- the feature is that the motion vector corresponding to the pixel of interest is determined based on the correlation determination processing.
- the motion vector determining unit determines a maximum pixel value and a minimum pixel value difference between the pixel value difference from the target pixel to which the motion vector is associated and the target pixel.
- One pixel is extracted as a feature pixel, and a motion vector corresponding to the pixel of interest is determined based on a correlation determination process based on the two feature pixels.
- the motion vector determination unit determines a motion vector corresponding to a target pixel based on a correlation determination process based on position information of the two characteristic pixels. Is determined.
- the motion vector determination unit determines a maximum pixel value and a minimum pixel value difference between a pixel value difference and a pixel of interest from a region near the pixel of interest with which the motion vector is associated.
- One pixel is extracted as a feature pixel, and a motion vector corresponding to the pixel of interest is determined based on a correlation determination process based on three pixels including the two feature pixels and the pixel of interest. .
- the motion vector determination unit extracts two feature pixels from a region near the pixel of interest to which the motion vector is associated, and It is characterized in that a motion vector corresponding to a target pixel is determined based on a correlation determination process based on pixel values of three pixels including the target pixel.
- a motion vector detection method for detecting a motion vector from moving image data comprising: an evaluation value table creation step of generating an evaluation value table based on pixel value correlation information between different frames on a time axis;
- a pixel correlation calculation step for calculating correlation information between different frames on the time axis based on a representative point matching process based on a representative point selected from one frame;
- a third aspect of the present invention provides:
- a computer program for detecting a motion vector from moving image data an evaluation value table creating step of generating an evaluation value table based on pixel value correlation information between different frames on a time axis,
- a pixel correlation calculation step for calculating correlation information between different frames on the time axis based on a representative point matching process based on a representative point selected from one frame;
- the operation is performed by applying at least one of the calculation result in the pixel correlation calculation step and a calculation result based on a pixel value difference value between the target pixel and pixels around the target pixel.
- the computer is characterized by having a program.
- the computer program according to the present invention is, for example, a storage medium or a communication medium provided in a computer-readable format to a computer system capable of executing various program code, such as a CD or FD. It is a computer program that can be provided by a recording medium such as an MO or a communication medium such as a network. By providing such a program in a computer-readable format, processing according to the program is realized on the computer system.
- system refers to a logical set of a plurality of devices, and is not limited to a device in which each component is located in the same housing.
- the correlation determination process is performed in consideration of not only the representative point but also the spatial waveform as the pixel level near the representative point.
- the weighting factor is determined based on the correlation information of the representative point pixel and the flag correlation information based on the flag data corresponding to the pixel value difference data of the target pixel and the pixel in the vicinity of the target pixel.
- W is calculated, and a reliability index is calculated as a calculated value based on the calculated weighting factor W and activity A as an index value indicating the complexity of the image data, and an evaluation value corresponding to the reliability index is calculated.
- the evaluation value table is generated by integrating With this configuration, it is possible to generate an evaluation value table based on the evaluation values weighted in consideration of the difference between the representative point and the pixel value near the representative point, and to generate a more accurate evaluation value table.
- a configuration is provided in which the flag data corresponding to the pixel value difference data of the target pixel and the pixel in the vicinity of the target pixel is calculated as a smaller number of bits than the pixel value difference data.
- the calculation of the flag correlation calculation process can be performed as a process with a small number of bits, and a highly accurate evaluation value table can be generated without reducing the processing speed, and the motion vector detection can be performed more accurately. It is possible to do. Therefore, by applying the present invention to an image processing device or the like that performs a coding process of moving image data, an efficient motion vector detection process can be performed, and the device can be downsized.
- the motion similarity between the representative point and the pixel near the representative point is determined. Is configured to judge that the reliability of the correlation determination result is high, calculate the reliability index, and generate the evaluation value table by integrating the evaluation values based on the reliability index ⁇ . A value table can be generated, and motion vector detection can be performed more accurately.
- the motion similarity between the representative point and the representative point in the vicinity of the representative point is detected, and the representative point and the neighboring representative point are detected on condition that the motion similarity is determined. It generates a reliability index i3 taking into account the pixel value difference from the point, and generates an evaluation value table in which evaluation values corresponding to the reliability index i3 are integrated.
- This configuration calculates the reliability index taking these parameters into account and generates the evaluation value table by integrating the evaluation values based on the reliability index, so that a more accurate evaluation value table can be generated. Motion vector detection can be performed more accurately.
- the flag correlation information based on the flag data corresponding to the pixel value difference data of the pixel of interest and the pixel in the vicinity of the pixel of interest is obtained.
- a weight coefficient W is calculated based on the calculated weight coefficient W and the image data.
- the motion vector corresponding to each pixel is selected and determined from a plurality of candidate vectors. Since feature pixels are extracted from the area in the vicinity of the target pixel to be associated, correlation determination is performed based on the position information or pixel value information of the feature pixel, and motion vector determination processing is performed, block matching processing is performed. Eliminates the need to apply It is possible to reduce both the evaluation value calculation such as the sum of absolute difference calculation as the calculation for correlation calculation, to improve the processing efficiency, and to reduce the memory for holding pixel values. And the hardware size can be reduced. Therefore, by applying the present invention to an image processing device or the like that executes encoding processing of moving image data, more accurate motion vector detection processing becomes possible.
- FIG. 1 is a diagram illustrating an outline of a block matching method.
- FIG. 2 is a diagram for explaining a problem of a motion vector detection process using representative point matching.
- FIG. 3 is a diagram illustrating an outline of a representative point matching method applied in a motion vector detection process.
- FIG. 9 is a diagram for describing an outline of a process.
- FIG. 5 is a diagram illustrating an outline of a motion vector determination process from a candidate vector in a motion vector detection process.
- FIG. 6 is a diagram showing a configuration of an embodiment of a motion vector detection device of the present invention that executes a motion vector detection process.
- FIG. 8 is a diagram showing a detailed configuration of an evaluation value table forming unit of the motion vector detection device of the present invention.
- FIG. 9 is a diagram illustrating a detailed configuration of a flag data calculation unit in an evaluation value table formation unit.
- FIG. 10 is a diagram illustrating a specific example of a process executed by a flag data calculation unit.
- FIG. 11 is a flowchart illustrating a specific example of a process executed by a weight calculation processing unit.
- FIG. 12 is a diagram showing a detailed configuration example 2 of an evaluation value table forming unit of the motion vector detection device of the present invention.
- FIG. 13 is a diagram illustrating a flowchart for explaining an evaluation value table formation processing sequence of the motion level detection device of the present invention.
- FIG. 14 is a diagram illustrating a representative point matching process.
- FIG. 15 is a diagram illustrating a process of performing similarity determination of a nearby representative point according to the present invention.
- FIG. 16 is a diagram showing a detailed configuration of an evaluation value table forming unit of the motion vector detection device of the present invention.
- FIG. 17 is a diagram illustrating processing of a motion similarity detection unit in an evaluation value table forming unit of the motion level detection device of the present invention.
- FIG. 18 is a diagram for explaining the processing of the weight calculation unit in the evaluation value table formation unit of the motion vector detection device of the present invention.
- FIG. 19 is a diagram illustrating a process of a weight calculation unit in the evaluation value table forming unit of the motion vector detection device of the present invention.
- FIG. 9 is a diagram illustrating a specific example of a reliability index j3 calculated by a unit.
- FIG. 21 is a diagram illustrating a process of a representative point stillness determination unit in the evaluation value table forming unit of the motion vector detection device of the present invention.
- FIG. 22 is a view illustrating a flowchart for explaining an evaluation value table formation processing sequence of the motion vector detection device of the present invention.
- FIG. 23 is a diagram showing an example of moving image data to be subjected to evaluation value table generation processing.
- FIG. 24 is a diagram showing an example of a valuation table generated by integrating data generated from the pixel correlation calculator and integrating the data.
- FIG. 25 is a diagram showing an example of an evaluation value table generated according to the processing of the evaluation value table forming unit according to the present invention.
- FIG. 26 is a diagram showing an example of an evaluation value table generated according to the processing of the evaluation value table forming unit according to the present invention.
- FIG. 27 is a diagram showing a detailed configuration of a motion vector determination unit of the motion vector detection device according to the present invention.
- FIG. 28 is a diagram illustrating processing in a temporary determination unit of a motion vector determination unit.
- FIG. 29 is a diagram showing a detailed configuration of an additional information calculation unit of the motion vector determination unit of the present invention.
- FIG. 30 is a diagram illustrating characteristic pixels detected by an additional information calculation unit.
- FIG. 31 is a diagram illustrating a feature pixel detection processing mode detected by an additional information calculation unit.
- FIG. 32 is a diagram illustrating details of processing executed in the motion vector determination unit of the motion vector detection device of the present invention.
- FIG. 33 is a flowchart illustrating a sequence of a motion vector determination process of the motion vector detection device of the present invention.
- FIG. 34 is a diagram illustrating a configuration example of an additional information calculation unit of the motion vector determination unit according to the present invention.
- FIG. 35 is a diagram showing a detailed configuration (Example 2) of a motion vector determination unit of the motion vector detection device of the present invention.
- FIG. 36 is a diagram showing a configuration example of an additional information calculation unit of the motion vector determination unit of the present invention.
- FIG. 37 is a diagram illustrating characteristic pixels detected by an additional information calculation unit.
- FIG. 38 is a diagram for explaining details of processing executed in the motion vector determination unit of the motion vector detection device of the present invention.
- FIG. 39 is a flowchart illustrating a sequence of a motion vector determination process of the motion vector detection device of the present invention.
- FIG. 40 is a diagram for explaining details of processing executed in the motion vector determination unit of the motion vector detection device of the present invention.
- FIG. 41 is a flowchart illustrating a sequence of a motion vector determination process of the motion vector detection device of the present invention.
- FIG. 42 is a flowchart illustrating a sequence of a motion vector determination process of the motion vector detection device of the present invention.
- one frame constituting moving image data is defined as one screen.
- An example in which a motion vector detection process in a frame is performed by a cross-verification process between screens (frames) will be described.
- the present invention is not limited to such a process example. It is also applicable to the case where a field is treated as one screen and motion vector detection processing is performed in units of one field.
- the processing example described below is mainly described as a processing example for moving image data based on a television signal.
- the present invention is applicable to various moving image data other than the television signal. It is also applicable. Further, even when a video signal is to be processed, either an interlace signal or a non-interlace signal may be used.
- a temporally continuous frame image forming a moving image for example, a current frame [F] 80 at time (t) and a previous frame [F] 70 at time (t ⁇ 1) shown in FIG. 3 are extracted. .
- the current frame [F] 80 is set as a reference frame
- the previous frame [F] 70 is set as m pixels X n t t-1
- the line is divided into a plurality of blocks, and a representative point Ry representing each block is set.
- the representative point of each block is, for example,
- Etc. are associated with pixel values representing the block.
- a corresponding search area 81 is set for the current frame [F] 80 in accordance with
- the pixel value of each pixel included in the cell 81 is compared with the pixel value of the representative point Ry71.
- the search area 81 is set, for example, as a pixel area of p pixels X q lines.
- the pixel value of any of the representative points of ac and the pixel value of each pixel in the search area 81 are compared and verified to calculate an evaluation value (for example, a frame difference or a determination result).
- the evaluation value is calculated for each shift (each pixel position) of the search area 81.
- Previous frame [F] Search area in current frame [F] 80 corresponding to each of the representative points of the block set in 70
- the search area corresponding to each representative point may be set so as to partially overlap an adjacent search area as shown in Fig. 4 (A).
- the search area is set to the area of pXq (pixel or line).
- the search area corresponds to the representative point 71a of the block set in the previous frame [F] 70ti.
- the search area 81b corresponding to the representative point 71b of the block set to 0 is set as an overlapping area.
- the search area is set in association with each representative point, and the comparison between each representative point and the pixels in the search area set corresponding to the representative point is executed. Based on the comparison value, for example, a higher evaluation value is set as the degree of correlation is higher (the higher the degree of matching of pixel values is), and an evaluation value corresponding to the constituent pixels of each search area is set.
- the evaluation values in each search area are integrated as shown in FIG. 4 (B), and as a result, an evaluation value table 90 as shown in FIG. 4 (C) is generated.
- the evaluation value table indicates the representative point Ryl t-i of each block set in, for example, n blocks set in the previous frame [F] 70
- It is calculated as an evaluation value based on comparison with each pixel in the search area, for example, as an integrated value of absolute difference values, and is formed as an evaluation value table having the same size as the search area.
- a peak occurs when the correlation between the pixel value and the representative point at each shift position (i, j) in the search area is high.
- the peak appearing in the evaluation value table corresponds to the movement of the displayed object on the screen of the moving image data.
- the evaluation value table having the same size as the search area (p X q) corresponds to the end point of the vector having that movement direction and distance.
- One peak appears at the position where If there are two differently moving objects in the screen (frame), an evaluation value table of the same size as the search area (p X q) In the bull, two peaks appear at two positions corresponding to the end points of two vectors having different movement directions and distances. If there is a stationary part, a peak corresponding to the stationary part also appears.
- a motion vector candidate (candidate vector) in the previous frame [F] 70 and the current frame [F] 80 as a reference frame is obtained.
- the most suitable candidate vector is selected from the extracted candidate vectors, and the candidate vector corresponding to each pixel is selected.
- the center pixel 91 indicates one pixel of the previous frame [F].
- This pixel has, for example, a luminance value (Ct).
- Ct luminance value
- pixel & 95, pixel 96, and pixel 097 are estimated as destinations from one pixel 91 of the previous frame ⁇ ] based on candidate vectors A, B, and C.
- the pixels of the current frame [F] at each pixel position are shown.
- the correlation with the pixel value of the block containing 1 is determined by the block matching process, the set with the highest correspondence is selected, and the candidate vector set for the selected set is set as the motion vector of the pixel 91.
- the reason for applying the block matching process is that the following problem occurs when the correspondence of only one pixel is detected.
- the MC residual with the pixel at the destination of the motion vector is considered to be the smallest, but in actuality, the pixel at the correct motion vector Is not necessarily the smallest among the candidate betatonores.
- the sum of absolute differences (SAD) is calculated according to the following equation as an index value of the correlation of a plurality of pixels included in the block specified by the candidate vector.
- F (x, y) is the luminance level of the frame of interest
- F (x + v, y + v) is the reference destination.
- the luminance level of the destination of the motion vector in the frame, M X N, is the block size used for evaluation.
- V is used as the motion vector of the pixel of interest.
- blocks such as block matching
- the representative point matching it is possible to set a representative point representing each block, calculate an evaluation value only for the set representative point, and set a candidate vector.
- the amount of calculation required for calculating the evaluation value can be reduced, and high-speed processing can be performed.
- FIG. 6 shows the configuration of an embodiment of the motion vector detection device of the present invention that executes the motion vector detection process, and the processing sequence of the motion vector detection process is shown in the flowchart of FIG.
- the motion vector detection device includes an evaluation value table forming unit 101, a candidate vector extracting unit 102, a motion vector determining unit 103, and a control unit (controller) 104.
- the evaluation value table forming unit 101 inputs image data to be subjected to motion vector detection processing via an input terminal, and generates an evaluation value table.
- the input image is, for example, image data obtained by a raster scan.
- the image data includes, for example, a luminance signal in a component signal of a digital color video signal.
- the input image data is supplied to the evaluation value table forming unit 101, and forms an evaluation value table having the same size as the search area based on the above-described representative point matching method. This is the process of step S101 in the flow of FIG.
- the candidate vector extraction unit 102 extracts a plurality of motion vectors as candidate vectors within one screen from the evaluation value table generated by the evaluation value table forming unit 101. That is, as described above, a plurality of candidate vectors are extracted based on the peaks appearing in the evaluation value table. This is the process of step S102 in the flow of FIG.
- the motion vector determination unit 103 calculates the correlation between pixels between frames associated with the candidate vector for each pixel of all the frames for the plurality of candidate vectors extracted by the candidate vector extraction unit 102. Judgment is performed by block matching or the like, and the candidate vector connecting the blocks having the highest correlation is set as the motion vector corresponding to the pixel. This is the process of step S103 in the flow of FIG. This process is the process described above with reference to FIG. [0090] The motion vector determination unit 103 performs an optimal vector selection process from the candidate vectors for all pixels included in one screen (frame), determines a motion vector corresponding to each pixel, Output the determined motion vector.
- a candidate connecting the block having the highest correlation by block matching processing is used.
- the vector is determined as the motion vector of the pixel of interest.
- the control unit 104 controls the processing timing in the evaluation value table forming unit 101, the candidate vector extraction unit 102, and the motion vector determining unit 103, stores intermediate data in a memory, and controls output processing.
- the above is the outline of the evaluation value table generation performed by the motion vector detection device based on the representative point matching and the motion vector detection processing to which the candidate vector method is applied.
- FIG. 8 shows the detailed configuration of the evaluation value table forming unit.
- the evaluation value table forming unit 200 in the motion vector detecting device of the present invention includes a pixel correlation operation unit 210, a flag data operation unit 220, a flag correlation operation unit 230, and comparison units 241 and 242. , A weight calculation processing unit 250, a calculation unit 260, and an evaluation value table calculation unit 270.
- the pixel correlation calculator 210 includes a representative point memory 211, a difference calculator 212 for calculating pixel value difference data, and an absolute value calculator 213 for calculating absolute value of difference data.
- the flag correlation calculation unit 230 includes a flag data storage memory 231 that stores the flag data calculated by the flag data calculation unit 220, a difference calculation unit 232 that calculates the difference data of the flag data, and an absolute value that calculates the absolute value of the difference data. It has a calculation unit 223.
- the weight calculation processing unit 250 includes an activity calculation unit 251, a weight calculation unit 252, a conversion tape insert (LUT) 253, and a decoder 254.
- the evaluation value table calculation unit 270 includes a calculation unit 260 An evaluation value composed of, for example, 8-bit data to be output is integrated in an evaluation value integrating unit 271 to generate an evaluation value table and store it in an evaluation value table memory 272.
- the pixel correlation calculator 210 performs a representative point matching process.
- Image data is input to the pixel correlation operation unit 210 via the input terminal, for example, in frame units.
- the image data input to the pixel correlation calculator 210 is supplied to the difference calculator 212 and the representative point memory 211.
- the image data stored in the representative point memory 211 is predetermined representative point data generated from image data input in units of frames, for example. For example, first look at Figure 3, Figure
- One representative point is set in a block set by dividing the screen described with reference to 4, for example, a block of m ⁇ n pixels.
- the representative points are
- Etc. is associated with pixel value data representing the block.
- image data pixel value data
- a signal from a control unit (controller) 104 see FIG. 6
- the representative point memory 211 is stored in the representative point memory 211 as representative point data.
- the representative point data of the previous frame read from the representative point memory 211 and the image data of the current frame are supplied to the difference calculator 212.
- the difference calculator 211 calculates a pixel value difference between the pixel value of the representative point of the previous frame and the pixel in the search area corresponding to each representative point set in the image data of the current frame, for example, the frame difference. (Correlation calculation result) is calculated and output to the absolute value calculation unit 213.
- the absolute value calculation unit 213 receives the representative point data of the previous frame input from the difference calculation unit 211 and the frame difference (correlation calculation result) based on the image data of the current frame, and calculates the frame difference absolute value. .
- the absolute value of the frame difference is input to the comparison unit 241 and compared with a predetermined threshold 1 (TH1).
- a predetermined threshold 1 TH1
- bit data for example, [1]
- bit data for example, [0]
- an evaluation value table has been generated as an integrated point using the output value from the comparison unit 241 as it is. That is, an evaluation value table is generated by accumulating the correlation operation results for all the representative points in one screen, and the extraction of candidate vectors is performed based on peaks (extreme values) appearing in the generated evaluation value table.
- the evaluation value table when the evaluation value table is generated by the correlation determination of only the representative point, a peak corresponding to an accurate vector is generated due to the erroneously detected integration point.
- the evaluation value table may not be generated.
- an evaluation value table is generated in consideration of the pixel level near the representative point, that is, the spatial waveform near the representative point.
- the 1-bit data indicating the presence or absence of strong correlation only in the correlation information of the representative point of the comparing unit 241 is converted into, for example, 8-bit data in the arithmetic unit 260 based on the output from the weight calculating unit 250, and the 8-bit data is
- the evaluation value is output to the evaluation value integration unit 271 to generate an evaluation value table based on the 8-bit data.
- the flag data calculation unit 220 performs a process of inputting image data and converting the relationship between adjacent pixel data into flag data for all input pixels.
- FIG. 9 shows a detailed configuration of the flag data calculation unit 220.
- the processing of the flag data calculation unit 220 will be described with reference to an example of FIG. 10 in which adjacent pixel difference values are converted into flags.
- Pixel signals as input signals are input in raster order, and supplied to the register 221 and the difference calculation unit 226 shown in FIG.
- the difference calculation unit 226 calculates a difference value between the output pixel level from the register 221 and the pixel level of a subsequent input pixel.
- the output from the register 221 is stored in the register 222, and is also supplied to the difference calculation unit 223, and the pixel level of the preceding pixel from the register 222 and the output from the register 221 are stored.
- the pixel level difference value of the succeeding pixel is calculated by the difference calculation unit 223.
- a difference value between the pixel level of (+1, y) 302 and the pixel (X-1, y) 303 is calculated.
- the data is output from each of the difference calculators 223 and 226 as 8-bit difference data.
- the conversion table 225 is configured by a table shown in FIG. 10 (b). Each quantizer 2
- the 24 and 227 refer to the conversion table 225 based on the value of the difference data d input from each of the difference calculators 223 and 226, and convert the data into predetermined flag data. For example,
- the quantizers 224 and 227 generate 4-bit flag data (0 to 15) based on the value of the 8-bit difference data d (0 to 255) input from each of the difference calculators 223 and 226. Then, the data is output to the flag data storage memory 231.
- the flag correlation operation unit 230 that performs the correlation judgment based on the flag data can perform the correlation judgment based on the data reduced to 4 bits, thereby speeding up the arithmetic processing and reducing the hardware scale. Is realized.
- the conversion process from the pixel difference data to the flag data is performed by the conversion process shown in FIG. Not limited to the processing using the table, for example, a configuration in which 8 bits of difference data of adjacent pixel data is reduced to 4 bits, or a method in which upper 4 bits of 8 bits of pixel value data are used as flag information may be applied.
- the above-described method of converting the adjacent pixel difference into a flag is an example of a method that can efficiently describe a spatial waveform with a small arithmetic circuit without using an extra line memory or the like because input pixels are input in raster order. is there. Also, since the difference data is converted to a flag, there is an advantage that rough waveform information composed of three pixels is not lost.
- the flag correlation calculation unit 230 stores the flag data input from the flag data calculation unit 220 in the flag data storage memory 231.
- the flag data of the previous frame and the flag data of the current frame read from the flag data storage memory 231 are supplied to the difference calculator 232.
- the difference calculation unit 232 calculates a flag difference between the flag data of the previous frame and the flag data of the current frame, and outputs the result to the absolute value calculation unit 233.
- the absolute value calculation unit 233 compares the flag data of the previous frame input from the difference calculation unit 232 with the flag data of the previous frame.
- the comparison unit 242 compares the absolute value of the flag data difference value input from the flag correlation operation unit 230 with
- the weight calculation processing unit 250 inputs the flux data difference value calculated by the flag correlation calculation unit 230 to the decoder 254, and calculates the activity of the image data.
- the activity is an exponent whose value increases as the image data becomes more complicated. In other words, in an edge area where the pixel value level such as the luminance level greatly changes in a narrow range, the activity is low in the case of image data having flat pixel values such as the sky where the activity is high. When the activity in the area near the representative point is large, it is determined whether or not the representative point has moved, and it has a great effect on human visual characteristics immediately.
- the peak in the evaluation value table depends on the area in the conventional method, and the size of the peak becomes small, so that it becomes difficult to extract candidate vectors.
- due to the characteristics of the image when pixel data with a large activity moves, a large level change occurs when the amount of movement is smaller than the sampling of the image data. Despite having such characteristics, when it is determined that the correlation based on the pixel level data is high, it can be determined that the reliability of the detected candidate motion vector is high.
- FIG. 11A shows a conversion table 253 held by the weight calculation processing unit 250.
- This is a table for performing an inverse conversion process of the conversion table described above with reference to FIG. 10 (B9), and is for obtaining a pixel value difference from the flag data. In some cases, activity is considered high.
- [0129] In the example shown in Fig. 11, [3] is input as flag data information of the left pixel, and [9] is input as flag data information of the right pixel.
- the decoder 254 obtains the corresponding representative value difference from the conversion table 253 shown in FIG. 11A and outputs it to the activity calculation unit 251.
- the representative value difference corresponding to the flag data information [3] of the left pixel is [1-110]
- the representative value difference corresponding to the flag data information [9] of the right pixel is [2]
- FIG. 11B shows a specific example of setting of pixel values when the above calculation is executed.
- Pixel value of target pixel 128,
- Pixel value of left adjacent pixel 118
- DR dynamic range
- the activity (A) calculated from the flag data has a value corresponding to the dynamic range (DR).
- the weight calculation processing unit 250 receives the flag data from the flag correlation calculation unit 230 and performs an activity based on the pixel value difference data converted based on the conversion table (see FIG. 11).
- the configuration may be such that activity A is calculated based on the activity.
- the activity calculation unit based on the flag data or the frame difference absolute value data
- the value of activity A calculated in 251 is output to weight calculating section 252.
- the weight calculation unit 252 calculates the weight W based on the correlation determination data as a result of comparison with the thresholds (TH1, TH2) input from the comparison units 241 and 242, and calculates the value of the activity A and the calculated weight. Calculate the reliability index based on W.
- the comparison unit 241 compares the pixel correlation calculation result calculated by the pixel correlation calculation unit with the threshold value (TH1) based on the pixel data of the representative point and the correlation in the search area set in a different frame.
- the comparison unit 242 compares the flag correlation calculation result calculated by the flag correlation calculation unit with the threshold value (TH2) based on the correlation between the flag data, and outputs the comparison result.
- the pixel data of the representative point and the flag data of the representative point are X, X, X, respectively, and the input pixel data and the flag data of the input pixel are ⁇ , ⁇ , ⁇ , respectively.
- thresholds TH1 and TH2 are fixed here, as another example, the thresholds may be calculated from image data and made variable.
- the weight calculation unit 252 calculates the degree of correlation as the weight based on the determination result of the above three equations (Equations 11 to 13). For example, when Equations 1, 2, and 3 are all satisfied, it can be determined that the correlation is high. On the other hand, when Equation 1 satisfies the condition but does not satisfy Equation 2 and Equation 3, it can be determined that the correlation is low. Therefore, the reliability of the detected candidate motion vector can be determined based on the degree of correlation determination.
- the degree of pixel-level correlation determination can also be calculated from the residual that can be calculated from Equation 1. For example, when the condition of Equation 1 is satisfied, assuming that a weighting factor W as an index indicating the degree of correlation is W,
- the weight calculation unit 252 calculates the reliability index H based on the weight coefficient W calculated by any of the methods described above and the activity A described above.
- the reliability index is It is calculated by an arithmetic unit.
- the weight calculation unit 252 outputs the reliability index calculated based on the weighting factor W and the activity A by the above equation (Equation 6) to the calculation unit 260, and is output from the comparison unit 241. Multiplication is performed on the result bit of the presence or absence of correlation based on the pixel at the representative point. As a result, when the result bit indicating the presence or absence of correlation based on the pixel at the representative point output from the comparison unit 241 is the bit [1] indicating the presence of the correlation, the value of the reliability index [H] is used as the evaluation value table calculation unit. Output to 270.
- the upper limit of the activity between the target pixel and the two adjacent pixels on the left and right is 6 bits
- the range of the above reliability is 0 ⁇ a ⁇ 252 which means that a weight of about 8 bits can be added to the evaluation value table compared to the conventional lbit, so that a highly reliable evaluation value table can be formed.
- the force described using the sum of absolute differences between two horizontal pixels adjacent to the target pixel as the activity may be a dynamic range or the like.
- the evaluation value table calculation unit 270 accumulates the evaluation values corresponding to the input reliability indices in the evaluation value accumulation unit 271 to generate an evaluation value table, and stores it in the evaluation value table memory 272.
- the evaluation value accumulating section 271 performs a process of accumulating the evaluation value corresponding to the reliability index input from the calculating section 260. As a result, an evaluation value table based on the reliability index is generated.
- step S201 a representative point is arranged (determined) in the previous frame image data.
- the representative point of each block is, for example,
- step S202 the current frame data is input. For example, they are input in raster order.
- step S203 the flag data calculation unit 220 (see FIG. 8) executes a process of converting the difference between the input pixel and the left and right adjacent pixels into a flag.
- this is a process of converting, for example, 8-bit pixel value difference data into 4-bit (015) flag data.
- step S204 a correlation determination process is performed between the representative point data and the input pixel data.
- This process is a process of the pixel correlation operation unit 210 and a process of the comparison unit 241 shown in FIG.
- the representative point data of the previous frame and the image data of the current frame read from the representative point memory 211 are supplied to the difference calculating unit 212, and are set to the pixel value of the representative point of the previous frame and the image data of the current frame.
- a pixel value difference with a pixel in the search area corresponding to each representative point to be calculated (for example, a frame difference calculation result) is output to the absolute value calculation unit 213, and the frame difference absolute value is calculated.
- the absolute value of the frame difference is input to the comparing section 241 and compared with a predetermined threshold 1 (TH1). If the absolute value of the frame difference is smaller than the threshold 1 (TH1), it is determined that there is a correlation, and the correlation is determined. Bit data (eg, [1]) indicating that there is no correlation is output from the comparing unit 241 and if the absolute value of the frame difference is equal to or greater than the threshold value 1 (TH1), it is determined that there is no correlation. ]) Is output from the comparison unit 241.
- TH1 threshold 1
- step S205 a correlation determination process is performed between the representative point flag data and the input pixel flag data. This process is executed by the flag correlation calculator 230 and the comparator 242 shown in FIG.
- the flag correlation calculator 230 stores the flag data input from the flag data calculator 220 in the flag data storage memory 231.
- the flag data of the previous frame and the flag data of the current frame read from the flag data storage memory 231 are supplied to the difference calculator 232.
- the difference calculation unit 232 calculates a flag difference between the flag data of the previous frame and the flag data of the current frame, and outputs the result to the absolute value calculation unit 233.
- the absolute value calculation unit 233 compares the flag data of the previous frame input from the difference calculation unit 232 with the flag data of the previous frame.
- the comparing unit 242 compares the absolute value of the flag data difference value input from the flag correlation calculating unit 230 with a predetermined threshold 2 (TH2), and determines the absolute value of the flag data difference value in advance. If smaller than threshold value 2 (TH2), there is correlation, and if the absolute value of the flag data difference value is equal to or larger than predetermined threshold value 2 (TH2), bit data without correlation is output to weight calculation section 252.
- TH2 predetermined threshold 2
- step S206 a reliability index is calculated.
- the process of calculating the reliability index is performed by the weight calculator 250.
- activity A is calculated in activity calculating section 251 based on flag data or frame difference absolute value data
- weight coefficient W is calculated based on outputs from comparing sections 241 and 242. Calculated based on activity A and weighting factor W,
- step S207 the reliability index ⁇ is output as an integrated point of the evaluation value table.
- step S208 it is determined whether or not the correlation determination process between all the representative points and the pixels in the search area has been completed, and if not completed, the process from step S202 is performed on the unprocessed pixels. When the processing for all the pixels is completed, the processing ends.
- the evaluation value calculation according to the present embodiment is performed, for example, with respect to one representative point 341 selected from the previous frame 340, for the pixel in the sub-cell 355 set in the current frame (input frame) 350.
- the search area 355 for the representative point 342 adjacent to the representative point 341 as well as performing the vector value comparison based on the presence or absence of the correlation by sequentially performing the pixel value comparison and calculating the correlation.
- the pixel value comparison is also performed for, the correlation is calculated, the vector is similarly detected based on the presence or absence of the correlation, and the motion vector detected by the correlation detection based on the plurality of representative points is calculated.
- an evaluation value with increased weight is set.
- the correlation determination is performed for a different representative point 342 near the representative point 341.
- the pixels are determined by the representative point 341 and the pixel 351 having high correlation.
- the motion vector V (vx, vy) 360 is determined to have high reliability, and is subjected to reliability addition processing to output a weighted evaluation value.
- the evaluation value table forming unit in the present embodiment performs one correlation determination for a certain representative point to determine an evaluation value
- the evaluation value table forming unit determines the motion level based on the correlation information of the representative point. The similarity with the motion vector based on the correlation information about the nearby representative point is determined. If there is a similar motion vector, a larger evaluation value is set for the reliability addition, and if there is no similar motion vector. Generates an evaluation value table by setting a small evaluation value.
- FIG. 16 shows a configuration of the evaluation value table forming unit 400 according to the present embodiment.
- the evaluation value table forming unit 400 in the motion vector detecting device of this embodiment It has a related calculation unit 410, a motion similarity detection unit 420, a comparison unit 430, a weight calculation unit 441, a calculation unit 442, a representative point stillness determination detection unit 450, and an evaluation value table calculation unit 270.
- the pixel correlation calculator 410 includes a representative point memory 411, a difference calculator 412 for calculating pixel value difference data, and an absolute value calculator 413 for calculating the absolute value of the difference data.
- the motion similarity detection unit 420 includes a pixel difference calculation unit 421 including a register (line memory) 422 and a difference calculation unit 423, a calculation unit 424, an absolute value calculation unit 425, and a correlation determination unit 426.
- the representative point stillness determination detection section 450 has a reference value memory 451, a comparison section 452, and a flag memory 453.
- the evaluation value table calculation unit 470 integrates the evaluation values, for example, composed of 8-bit data output from the calculation unit 442, into the evaluation value integration unit 471, generates an evaluation value table, and stores it in the evaluation value table memory 472. I do.
- the pixel correlation calculator 410 performs a representative point matching process.
- Image data is input to the pixel correlation operation unit 410 via the input terminal, for example, in frame units.
- the image data input to the pixel correlation calculator 410 is supplied to the difference calculator 412 and the representative point memory 411.
- the image data stored in the representative point memory 411 is, for example, predetermined representative point data generated from image data input in frame units.
- one representative point is set in a block set by dividing the screen described above with reference to FIGS. 3 and 4, for example, a block of m ⁇ n pixels.
- the representative points are
- Etc. is associated with pixel value data representing the block.
- image data (pixel value data) at pixel positions spatially evenly decimated from an input frame image is transmitted from the control unit (controller) 104 (see FIG. 6).
- the signal is selected by signal timing control and stored in the representative point memory 411 as representative point data.
- the difference calculation unit 411 calculates a pixel value difference between the pixel value of the representative point of the previous frame and the pixel in the search area corresponding to each representative point set in the image data of the current frame, for example, the frame difference. (Correlation calculation result) is calculated and output to the absolute value calculation unit 413.
- the absolute value calculation unit 413 receives the representative point data of the previous frame input from the difference calculation unit 411 and the frame difference (correlation calculation result) based on the image data of the current frame, and calculates the frame difference absolute value. .
- the absolute value of the frame difference is input to the comparing section 430 and compared with a predetermined threshold (TH). If the absolute value of the frame difference is smaller than the threshold (TH), it is determined that there is a correlation, and the presence of the correlation is determined. If the bit data (eg, [1]) indicating that the frame difference absolute value is greater than or equal to the threshold (TH) is output from the comparing unit 430, it is determined that there is no correlation, and the bit data indicating that there is no correlation (eg, [0]) Are output from the comparing section 430.
- TH predetermined threshold
- an evaluation value table has been generated as an integration point using the output value from the comparing section 430 as it is. That is, an evaluation value table is generated by integrating the correlation operation results for all the representative points in one screen, and the extraction of candidate vectors is performed by peaks (extreme values) appearing in the generated evaluation value table.
- peaks extreme values
- the evaluation value table forming unit in the motion vector detecting apparatus performs one correlation determination for a certain representative point to determine an evaluation value. Determines the similarity between the motion vector based on the correlation information of the point and the motion vector based on the correlation information about the nearby representative points, and if there is a similar motion vector, sets a larger evaluation value for reliability addition. If there is no similar motion level, an evaluation value table is generated by setting a smaller weight and an evaluation value.
- the power S for executing these processes is the motion similarity detection unit 420, representative point stillness determination unit 450, weight calculation unit 441, and calculation unit 442 shown in FIG. The processing of each of these units will be described below.
- the motion similarity detection unit 420 receives the image data and performs a pixel difference calculation unit 42
- the input pixel data is stored in the register (line memory) 422 in 1.
- the input frame 510 shown in FIG. 17 is stored in the register (line memory) 422.
- XI and X5 indicate the pixel values of the pixels 511 and 512, respectively.
- the interval between the pixel 511 and the pixel 512 is set equal to the representative point interval dx set in the previous frame 500.
- the difference value D is input to the calculation unit 424, and the calculation result of the calculation unit is used as the absolute value calculation unit.
- step 2 [S2 shown in FIG. 17 based on the input value from the pixel difference operation unit 421 and the pixel value of the representative point pixel of the previous frame from the representative point memory 411. ]
- step 3 [S3] are executed.
- step 2 [S2] the pixel value Q of the representative point pixel 502 of the previous frame 500 input from the representative point memory 411 shown in FIG. IQ—Calculate X5 I.
- step 3 [S3], the pixel value P of the representative point pixel 501 of the previous frame 500 input from the representative point memory 411 shown in Fig. 6 and the difference absolute value I of the pixel 511 of the input frame 510 are obtained. Calculate P-X5 + DI.
- FIG. 17 shows an example of detecting the similarity of the motion vector between the focused representative point 501 and the neighboring representative point 502 on the right side.
- the correlation between one input pixel and multiple representative points is determined. Therefore, IQ—X5 I and IP—XI I are not calculated in parallel. Instead of calculating IP—XI I, -X5 + DI is calculated. Therefore, as in the calculation procedure S1 S3 shown in FIG. 17, before performing representative point matching in advance, the difference value D between the input pixel in the input frame and the pixel separated by the representative point interval is calculated. are doing.
- the correlation determination unit 426 performs the process of step 4 [S4] shown in Fig. 17, that is, the comparison process between the predetermined threshold value [TH] and the pixel difference value obtained in steps 2 and 3,
- the weight calculator 441 can trust the motion vector obtained by correlation detection corresponding to a nearby representative point if the motion vector is similar. Then, a reliability index is calculated based on the reliability, and the reliability index is output to the calculation unit 442.
- the operation unit 442 inputs the reliability index and outputs the output from the comparison unit 430, that is, the representative point and each pixel of the search area as a result of the representative point matching executed by the pixel correlation operation unit 410. As a result of the presence or absence of the correlation based on the correlation result, enter bit [1] if there is correlation, and input bit [0] if there is no correlation.
- arithmetic unit 442 calculates the bit 1 based on the determination that there is a similar motion of the representative point near the representative point input from the weight calculation unit with respect to bit 1.
- the multiplication or addition of the reliability index i3 is performed, and XI or +1 is output to the evaluation value integrating unit 471 as the final evaluation.
- the weight calculating section 441 calculates the reliability index / 3 when it is determined that the output from the motion similarity detecting section 420 has similarity. The details of the calculation process of the reliability index / 3 in the weight calculation unit 441 will be described.
- the motion vector of the representative point of interest and the neighboring representative point are similar, the spatial gradient between the representative points, that is, a large pixel between the representative points as shown in FIG. If there is a value difference, an area having a brightness level difference in the image will move, and the motion vector Judge that the reliability is higher. On the other hand, if there is no large pixel value difference between the representative points as shown in Fig. 18 (b), this indicates that the area with a small difference in luminance level moves in the image, for example, the background area and the sky area In such cases, the reliability of the motion vector is determined to be low even when similarity is detected.
- the weight calculating section 441 calculates the reliability index based on the magnitude of the spatial gradient 550 between the representative points.
- the weight calculation unit 441 receives the determination that there is motion similarity between the representative points from the motion similarity detection unit 420, the weight level of the focused representative point of interest is determined to be Pm, and there is similar motion.
- Pn be the luminance level of the N neighboring representative points
- index / 3 representing the reliability of the motion vector is given by the following equation:
- FIG. 20 shows a specific example of the reliability index ⁇ calculated according to the above equation.
- FIG. 20 shows an example in which four representative points around the focused representative point 550 are selected as neighboring representative points, and the similarity of each motion is detected by the motion similarity detection unit 420.
- the arrangement of the representative points is not a fixed arrangement but is set as a variable arrangement by calculating a certain parameter from image data
- the distance from the representative point of interest also increases. It is preferable to set the reliability index in consideration of the configuration. That is, it is determined that the closer the distance between the representative points is, the higher the reliability of the similar motion vector is, and the smaller the distance between the representative points is, the higher the reliability index is set.
- the reliability index i3 of the motion vector when the motion vector of the focused representative point and the neighboring representative point are similar is calculated as a value reflecting at least one of the following parameters.
- the weight calculation unit 441 receives from the motion similarity detection unit 420 the similarity determination results of N neighboring representative points corresponding to a specific representative point, and further, as shown in FIG.
- the spatial gradient (pixel value difference information) between the representative points is obtained based on the output from the unit 410, the reliability index i3 is calculated according to the above equation, and output to the arithmetic unit 442.
- the determination whether or not the pixel position of the representative point is in the still area is performed by the representative point stillness determination unit 450.
- the representative point stillness determination unit 450 includes a reference value memory 451, a comparison unit 452, and a flag memory 453.
- the comparison unit 452 receives the result of the representative point matching process executed in the pixel correlation calculation unit 410, and compares the result with a preset reference value stored in the reference value memory 451. It is determined whether the point is in the stationary area.
- the stationary point of the representative point Ry of the frame [F] is determined by the representative point Ry of the previous frame [F] and the current frame.
- Judgment is made based on the frame difference between the representative point Rt-it-iy of the previous frame [F] and the point Py of the previous frame [F], instead of using the frame difference from the point Qy of the frame [F]. Because Ry and Q
- the stationary point of the representative point Ry can be determined only when the Qy data of the current frame is supplied, but when using the frame difference between Ry and Py, the previous frame is used. When all the data of [F] is supplied, the representative that exists in the previous frame [F]
- the pixel difference absolute value is calculated by the pixel correlation calculator 410 at the representative point Py of the frame [F].
- the comparison unit 452 calculates the representative point Ry of the previous frame [F] calculated by the pixel correlation operation unit 410 with the representative point Ry.
- the pair value is smaller than the preset reference value stored in the reference value memory 451, it is assumed that the representative point Ry of the previous frame is in the still area, and the still flag (0) is set in the flag memory.
- the reliability index ⁇ calculated according to the above equation is output to the calculator 442.
- Arithmetic unit 442 performs an operation of adding or multiplying the reliability index / 3 to the output from comparison unit 430, or performing an operation process of setting the result in consideration of the reliability index as an output evaluation value, to thereby obtain an evaluation value integrating unit. Output to 471.
- the evaluation value table calculation unit 270 outputs the evaluation value reflecting the input reliability index ⁇ to the evaluation value integration unit 271 to generate an evaluation value table, and stores it in the evaluation value table memory 272. Store.
- Evaluation value integration section 271 performs processing of integrating evaluation values reflecting reliability index ⁇ input from operation section 260. As a result, an evaluation value table based on the reliability index is generated.
- step S301 a representative point is arranged (decided) in the previous frame image data.
- the representative point of each block is, for example,
- Etc. are associated with pixel values representing the block.
- step S302 the current frame data is input. For example, they are input in raster order.
- step S304 a correlation determination process is performed between the representative point data and the input pixel data.
- This processing is the processing of the pixel correlation operation unit 410 and the processing of the comparison unit 430 shown in FIG.
- the representative point data of the previous frame and the image data of the current frame read from the representative point memory 411 are supplied to the difference calculation unit 412, and are set to the pixel value of the representative point of the previous frame and the image data of the current frame.
- a pixel value difference from a pixel in the search area corresponding to each representative point (for example, a frame difference calculation result) is calculated and output to the absolute value calculation unit 413, and the frame difference absolute value is calculated.
- the absolute value of the frame difference is input to the comparing section 430 and compared with a predetermined threshold value (TH). If the absolute value of the frame difference is smaller than the threshold value (TH), it is determined that there is a correlation, and bit data indicating the presence of the correlation is determined. (Eg, [1]) is output from the comparison unit 430, and when the absolute value of the frame difference is equal to or larger than the threshold (TH), it is determined that there is no correlation, and bit data indicating no correlation (eg, [0]) is compared. Output from unit 430.
- step S305 the motion similarity of the nearby representative point is detected from the difference data between the input pixel and the representative point.
- This process is a process of the motion similarity detection unit 420. The process is performed by the motion similarity detection unit 420.
- the attention representative point detected based on the correspondence between the focused representative point in the previous frame and the highly correlated pixel in the input frame corresponding to the neighboring representative point. It is determined whether or not the motion of the nearby representative points is similar.
- the motion similarity detection unit 420 outputs a similarity determination result to the weight calculation unit 441. It should be noted that the determination result of the presence or absence of similarity is executed for a plurality of nearby representative points of the noted representative point, and the respective result data is output to the weight calculating unit 441.
- step S306 a reliability index ⁇ is calculated.
- the calculation process of the reliability index / 3 is executed by the weight calculation unit 441.
- the reliability index of the motion vector when the motion vector of the focused representative point is similar to that of the neighboring representative point is calculated as a value reflecting at least the following parameters.
- step S307 the reliability index ⁇ is output as an integrated point of the evaluation value table.
- step S308 it is determined whether or not the correlation determination processing between all the representative points and the pixels in the search area has been completed. If not, the processing from step S302 onward is performed on the unprocessed pixels. When all the pixels have been processed, the process ends.
- the reliability index ⁇ is calculated by combining the reliability index a described above and the reliability index ⁇ described above, and
- the configuration of the evaluation value table forming unit is a configuration having both the configuration shown in FIG. 8 and the configuration shown in FIG.
- the evaluation value table is configured by adding lbit (correlation / non-correlation) obtained as a result of determining the presence or absence of correlation only in representative point matching.
- FIG. 24 shows the value table.
- the evaluation value table shown in FIG. 24 is a two-dimensionally expressed evaluation value table showing only peaks corresponding to motion vectors in the vertical direction (one Y or + Y direction).
- FIG. 25 shows two-dimensional data of an evaluation value table generated based on the reliability index: a described above, that is, the reliability index ⁇ based on the activity A and the weight coefficient W.
- this evaluation value table not only peaks corresponding to still pixels but also peaks corresponding to the object D moving in the vertical direction ( ⁇ direction) appear.
- the occupied area in the display area of the image data is small, and the peak corresponding to the movement of the object is calculated. It can appear in the evaluation value table, and enables accurate candidate vector extraction and motion vector determination processing.
- Fig. 26 shows a reliability index: ⁇ , that is, a reliability index calculated in consideration of motion similarity detection between a focused representative point and a nearby representative point and a spatial gradient (pixel value difference) between the representative points.
- ⁇ a reliability index calculated in consideration of motion similarity detection between a focused representative point and a nearby representative point and a spatial gradient (pixel value difference) between the representative points.
- the evaluation value tape generated based on the reliability index ⁇ calculated in consideration of the detection of the motion similarity between the focused representative point and the neighboring representative point and the spatial gradient (pixel value difference) between the representative points.
- the peak corresponding to the movement of an object with a small occupation area can appear in the evaluation value table, although the activity ⁇ of the representative point in the display area of the image data is small. Can be determined.
- the motion vector detection device to which the representative point matching is applied has an evaluation value table forming unit 101, a candidate vector extracting unit 102, and a motion vector determining unit 103, and A value table forming unit 101 generates an evaluation value table based on the input image data, a candidate vector extracting unit 102 extracts a plurality of candidate vectors from the evaluation value table, and further a motion vector determining unit 103 Is performed to determine a motion vector corresponding to each pixel from the candidate vectors.
- a motion vector determination process based on feature pixel position information is executed without applying block matching to the motion vector determination process executed in the motion vector determination unit 103.
- the details of the processing of the motion vector determination unit will be described below.
- FIG. 27 shows the detailed configuration of the motion vector determination unit of the motion vector detection device according to the present embodiment.
- the motion vector determination unit 1200 shown in FIG. 27 corresponds to the motion vector determination unit 103 of the motion vector detection device shown in FIG.
- the motion vector determination unit 1200 shown in FIG. 27 inputs a plurality of pieces of candidate vector information determined based on the above-described evaluation value table from the candidate vector extraction unit 102 shown in FIG. A process for determining a motion vector to be associated is executed.
- the motion vector determination unit 1200 includes a pixel correlation determination unit 1210, a tentative determination unit 1221, a gate 1222, an additional information calculation unit 1250, and a characteristic pixel position correlation determination unit 1230.
- the pixel correlation determination unit 1210 has a current frame memory 1211, a previous frame memory 1212, a pixel value difference calculation unit 1213, and an absolute value calculation unit 1214.
- the characteristic pixel position correlation determination unit 1230 includes a current frame-added information memory 1231, a previous frame additional information memory 1232, and a correlation determination unit 1233. Details of the additional information calculation unit 1250 will be described with reference to FIG.
- the pixel correlation determination unit 1210 receives an image signal.
- This input image is, for example, image data obtained by raster scanning.
- the image data includes, for example, a luminance signal in a component signal of a digital color video signal.
- Image data is input for each frame data, firstly stored in the current frame memory 1211 and then stored in the previous frame memory 1212. Therefore, the pixel correlation determination unit 1210 holds two consecutive frame data in the memory.
- pixel correlation determining section 1210 receives candidate vector information from candidate vector extracting section 102 shown in FIG. 6, and should determine each pixel of the previous frame, that is, a motion vector, based on the candidate vector information. For each pixel (pixel of interest), the pixel position of the current frame indicated by the plurality of candidate vectors is specified, and the difference between the pixel of interest of the previous frame and the pixel of the current frame indicated by the plurality of candidate vectors is calculated as the pixel value difference. Calculated by the calculation unit 1213 and calculated as an absolute value The output unit 1214 calculates the absolute value of the difference, and outputs the calculated absolute value to the temporary determination unit 1221.
- FIG. 28 shows input frame (current frame) image data 1300 and past frame (previous frame) image data 1310. These frame data are data stored in the current frame memory 1211 and the previous frame memory 1212.
- the pixel correlation determination unit 1210 receives candidate vector information from the candidate vector extraction unit 102 shown in FIG. 6, and is a candidate vector of these forces a and g shown in FIG. These are the candidate vectors obtained by the peak detection in the evaluation value table described above.
- the pixel correlation determination unit 1210 determines, for each pixel in the previous frame based on the candidate vector information, that is, for each pixel (pixel of interest) for which a motion vector is to be determined, the current frame indicated by a plurality of candidate vectors. Is specified. For example, for each pixel of interest 1311 of the past frame (previous frame) image data 1310 in FIG. 28, the pixel positions A to G of the current frame indicated by a plurality of candidate vectors are specified.
- the pixel value difference calculation unit 1213 calculates a difference between the pixel of interest 1311 of the past frame (previous frame) image data 1310 and each of the pixel values of the pixel positions A to G of the current frame indicated by the candidate vector.
- the absolute value calculation unit 1214 calculates these difference absolute values, that is, the pixel difference absolute values (MC residuals), and outputs them to the tentative determination unit 1221.
- the temporary determination unit 1221 narrows down the motion vector corresponding to the pixel 1311 of interest from the n absolute difference values (MC residual values) [di]. Specifically, from n ((absolute difference (MC residual) [di]), the one with the smaller MC residual is selected.
- the selected candidate vector information is output from the tentative determination unit 1221 to the gate 1222, and from the candidate vector information input from the candidate vector extraction unit 102 shown in FIG. Only the force is input to the feature pixel position correlation determination unit 1230.
- the characteristic pixel position correlation determining unit 1230 executes a process of determining a motion vector corresponding to one pixel from the candidate vectors selected by the tentative determining unit 1221.
- the feature pixel position correlation determination unit 1230 performs a pixel value correlation determination process based on the feature pixel position information of the neighboring pixels of the pixel of interest input from the additional information calculation unit 1250, and the provisional determination unit 1221 A process of determining one motion vector from the selected candidate vectors is performed.
- the additional information calculation unit 1250 extracts, for example, a pixel having the largest absolute value of a pixel value difference from an adjacent pixel among pixels near the pixel of interest as a feature pixel.
- the additional information calculation unit 1250 shown in FIG. 29 includes an adjacent pixel difference absolute value calculation unit 1251 including a register 1252, a pixel value difference calculation unit 1253, and an absolute value calculation unit 1254, a register 1255, and a maximum difference value detection unit 1256. , Having a maximum difference pixel position detecting unit 1257 and a register 1258
- the adjacent pixel difference absolute value calculation unit 1251 inputs an image signal to be subjected to motion vector detection processing, and stores image frame data in the register 1252.
- the pixel value difference calculation unit 1253 sequentially calculates the difference between the adjacent pixels of the image data stored in the register 1252, and the absolute value calculation unit 1254 sequentially calculates the absolute value of the difference between the adjacent pixels to obtain the maximum difference value. Output to detector 1256.
- the additional information calculation unit 1250 sequentially calculates the absolute value of the difference between the pixels adjacent to the image data stored in the register 1252, so that, for example, in the pixels in the vicinity of the pixel of interest, the absolute value of the difference from the P-contact pixel is the smallest. Identify large pixels.
- the pixels in the vicinity of the pixel of interest are pixel areas that are set in advance, such as a range of 8 + 7 pixels before and after the pixel of interest in the horizontal direction.
- the pixel 1282 with the largest absolute difference from the pixel in contact with the pixel of interest is specified.
- the maximum difference value detection unit 1256 shown in FIG. 29 executes a comparison process while storing the difference absolute value of the sequentially input adjacent pixel in the register 1255, and executes a predetermined area, that is, -8— + 7 of 16 A pixel difference value having a maximum pixel value difference absolute value for each pixel is detected. That is, as shown in FIG. 31, the pixel difference values MAXa, MAXb, and MAXc having the maximum pixel value difference absolute value every 16 pixels are sequentially detected.
- the pixel information having the absolute value of the maximum pixel difference for each predetermined region (for example, for every 16 pixels) detected by the maximum difference value detection unit 1256 is input to the maximum difference pixel position detection unit 1257, and for each predetermined region (The maximum difference pixel position information (for example, every 16 pixels) is detected and stored in the register 1258.
- the pixel position of the pixel 1282 at which the absolute value of the difference from the adjacent pixel is maximum is: + 4] pixel position.
- the maximum difference pixel position information for each predetermined area (for example, for every 16 pixels) stored in the register 1258 is sequentially output to the current frame additional information storage memory 1231.
- the characteristic pixel position information from the additional information calculation unit 1250 described with reference to FIG. 29 is stored in the current frame-attached information memory 1231 of the characteristic pixel position correlation determination unit 1230 shown in FIG.
- the current frame-attached caro information memory 1231 stores the maximum difference pixel position information for each predetermined area (for example, every 16 pixels) extracted corresponding to one frame.
- the information in the current frame additional information memory 1231 is moved to and stored in the previous frame additional information memory 1232 for each frame processing step.
- the current frame additional information memory 1231 is further input with the output from the tentative determination unit 1221, ie, the selection candidate vector information selected by the tentative determination unit 1221.
- the correlation determination unit 1233 determines the only motion vector corresponding to the target pixel from the selection candidate vectors selected by the tentative determination unit 1221 based on the feature pixels extracted in the vicinity region of the target pixel. Execute the processing to be performed.
- FIG. 32 shows an input frame (current frame) similar to that described above with reference to FIG. Frame) image data 1301 and past frame (previous frame) image data 1310.
- the input frame (current frame) image data 1300 shows candidate vectors a, c, and e selected by the candidate vector selection processing by the tentative determination unit described above with reference to FIG.
- the correlation determination unit 1233 in the characteristic pixel position correlation determination unit 1230 includes a pixel of interest selected from the past frame (previous frame) image data 1310, that is, a region in the vicinity of the pixel of interest 1311 as a pixel associated with the motion vector. Based on the characteristic pixel position data in, one motion vector is determined from the candidate vectors a, c, and e.
- the feature pixel position in the vicinity area of the pixel of interest 1311 is the pixel position [13]. It is determined based on input information from the arithmetic unit 1250.
- the neighboring region is shown as an eight-pixel region of (141 + 3) pixels of interest.
- the correlation determining unit 1233 calculates the correlation between the characteristic pixel position in the vicinity of the pixel of interest 1311 and the characteristic pixel position in the vicinity of the corresponding pixel position A, C, E indicated by the selection candidate vectors a, c, e. Is determined.
- the characteristic pixel appears at the position of [-3].
- This feature pixel is a feature pixel selected for each pixel area in the predetermined area unit in the additional information calculation unit described above, and is the maximum pixel of the adjacent pixel difference absolute value in the area set as the area near the pixel position A. is there.
- the characteristic pixel appears at the position [13] when the pixel of interest position is [0].
- This feature pixel position corresponds to the feature pixel position [13] in the neighborhood 1301 of the pixel A indicated by the candidate vector a . It does not correspond to the characteristic pixel positions [1-2] and [+2] in the neighborhoods 1302 and 1303 of the pixels C and E indicated by the other two candidate vectors c and e .
- the candidate vector a is selected and determined as the motion vector corresponding to the pixel of interest 1311.
- the correlation determination unit 1233 sequentially sets such processing as the pixel of interest for the constituent pixels of the frame, and executes the correlation determination processing of the characteristic pixel position of the neighboring region in the same manner as described above for each set pixel of interest. Then, a motion vector corresponding to each pixel constituting the frame is selected and determined from the selection candidate vectors.
- the motion vector determining unit shown in FIG. 27 specifies the characteristic pixel position based on the pixel value of the region near the target pixel, and determines the position of the characteristic pixel in the region adjacent to the target pixel.
- the correlation between the corresponding relative position and the characteristic pixel position in the vicinity of the pixel specified by the candidate vector selected by the tentative determination unit 1221 is determined, and the candidate specifying the pixel having the highest correlation characteristic pixel position is determined.
- the vector is determined as the motion vector corresponding to the pixel of interest.
- the constituent pixels of the frame image are sequentially set as the pixel of interest, and FIG. Is executed for each pixel of interest.
- step S1201 is a process of calculating the absolute value of the difference between the pixel of interest and the pixels specified by n candidate vectors, that is, the MC residual dn.
- This processing is executed in the pixel correlation determination unit 1210 of the motion vector determination unit 1200 in FIG.
- Steps S1202 and S1204 are processes in the provisional determination unit 1221.
- the tentative judgment unit 1221 inputs n difference absolute values (MC residuals) [di] and detects the minimum value of the n MC residuals: dmin.
- step S1204 is a tentative determination process executed as a process of selecting from candidate vectors described above with reference to FIG. If it is determined that di-dmin ⁇ TH is not satisfied, it is determined that the vector is not a selection candidate vector, and the correlation determination based on the characteristic pixel position of the neighboring region of the target pixel performed in steps S1205 and S1206 is performed. Proceed to step S1207.
- step S1204 if didmin ⁇ TH is satisfied, the selection is made as a selection candidate vector, and the flow advances to step S1205.
- the processing of steps S1205 and S1206 is the processing of the characteristic pixel position correlation determination section 1230 of the motion vector determination section 1200 shown in FIG.
- step S1205 the correlation between characteristic pixel positions is determined. As described above with reference to FIG. 32, this processing is based on the correlation between the position of the characteristic pixel in the area near the pixel of interest and the characteristic pixel position in the area near the pixel selected as the selection candidate vector, that is, the position. This is a process of determining whether or not there is a correspondence.
- step S1206 a process of selecting a candidate having a higher correlation is executed.
- step S1207 it is determined whether or not all candidate vectors have been verified based on the value of i. In this case, the process proceeds to step S1209, where the value of i is updated (incremented by one), and the processing of step S1204 and thereafter is repeatedly executed.
- step S1208 the process proceeds to step S1208, at which point the position of the feature pixel in the vicinity of the target pixel calculated in step S1205, that is, the region having the highest correlation, is calculated. Then, a candidate vector corresponding to a pixel having a high correlation between characteristic pixel positions in the vicinity of a pixel selected as a selection candidate vector is determined as a motion vector corresponding to a target pixel to be processed.
- the pixel of interest and the neighboring characteristic pixel positions that are not subjected to block matching processing are determined. This is a configuration using information. Therefore, it is necessary to perform a correlation calculation based on a large number of pixel values according to the size of a block to be applied in the block matching process, and efficient processing can be performed.
- the target pixel position is set to [0], and 16 pixels of _8 + 7 and 8 pixels of 14 + 3 are described as examples. These can be set as arbitrary areas.
- the neighboring area is set only on the right side in the horizontal direction, adjacent pixel comparison processing based on the input pixel values in the raster scan order can be performed.
- the value memory area can be reduced, and the hardware scale can be further reduced.
- a neighboring area of a pixel is set, and the additional information calculation unit 1250 calculates, as characteristic pixel information, a pixel position where the absolute value of the neighboring pixel difference in the neighboring area is the maximum, and calculates the characteristic pixel information.
- a configuration for applying information to correlation determination that is,
- Fig. 34 shows an example of the configuration of the additional information calculation unit 1250 in the case where (b) the position of the pixel having the maximum absolute value of the adjacent pixel difference and the pixel value are applied as the correlation determination information.
- the configuration of the additional information calculation unit 1250 shown in FIG. 34 is almost the same as the configuration of the additional information calculation unit 1250 shown in FIG. 29, and includes a register 1252, a pixel value difference calculation unit 1253, and an absolute value calculation unit 1254.
- An adjacent pixel difference absolute value calculation unit 1251 inputs an image signal to be subjected to motion vector detection processing, and stores image frame data in the register 1252.
- the pixel value difference calculator 1253 sequentially calculates the difference between adjacent pixels of the image data stored in the register 1252, and the absolute value calculator 1254 sequentially calculates the absolute value of the difference between Output to value detector 1256.
- the maximum difference value detection unit 1256 executes a comparison process while storing the difference absolute value of the sequentially input adjacent pixel in the register 1255, and performs the maximum pixel value difference absolute value in a predetermined area, for example, every 16 pixels. Is detected.
- the pixel information having the absolute value of the pixel value difference for each predetermined area (for example, every 16 pixels) detected by the maximum difference value detection unit 1256 is input to the maximum difference pixel position detection unit 1257, and for each predetermined area (for example, 16 pixels).
- the maximum difference pixel position information (for each pixel) is detected and stored in the register 1258.
- the additional information calculation unit 1250 described with reference to Fig. 29 above outputs only the position information stored in the register 1258 to the current frame-attached information memory 1231, and performs a correlation determination process based on the position information. It was configured to perform.
- the additional information calculation unit 1250 shown in FIG. 34 stores not only the position information indicating the maximum difference pixel position stored in the register 1258 but also the pixel value information of the maximum difference pixel position from the register 1255 into the current frame attached information of the current frame. Output to memory 1231.
- the characteristic pixel position correlation determination unit 1230 shown in Fig. 27 applies the correlation between the two data, that is, the two pieces of information, that is, the position information of the pixel having the maximum absolute value of the adjacent pixel difference and the pixel value information. Execute the judgment process.
- the two-point matching described above uses the assumption that an object moves with an area. For example, if a feature pixel is set at a position far from the pixel of interest, there is a possibility that the object is the same object. descend. In view of this, a configuration may be adopted in which processing is performed in which the distance to the feature pixel that is not only the degree of coincidence of the feature pixel position is reflected in the correlation. In other words, the feature pixel at a position closer to the target pixel is given a higher weight at the time of correlation determination, and the feature pixel at a position farther from the target pixel is given a smaller weight at the time of correlation determination.
- FIG. 35 shows a detailed configuration of the motion vector determination unit of the motion vector detection device according to the present embodiment.
- the motion vector determination unit 400 shown in FIG. 35 corresponds to the motion vector determination unit 103 of the motion vector detection device shown in FIG.
- the motion vector determination unit 400 shown in FIG. 35 inputs a plurality of pieces of candidate vector information determined based on the above-described evaluation value table from the candidate vector extraction unit 102 shown in FIG. A process of determining a motion vector to be associated with each motion is executed.
- the motion vector determination unit 400 includes a pixel correlation determination unit 1410, a provisional determination unit 1421, an additional information calculation unit 1450, a current frame additional information memory 1422, and a previous frame addition.
- An information memory 1423 and a neighboring area information correlation determination unit 1430 are provided.
- the pixel correlation determination unit 1410 includes a current frame memory 1411, a previous frame memory 1412, a pixel value difference calculation unit 1413, and an absolute value calculation unit 1414.
- the neighboring area information correlation determining unit 1430 has registers 1431 and 1432, a gate 1433, and a correlation determining unit 1434. Details of the additional information calculation unit 1450 will be described with reference to FIG.
- the pixel correlation determination unit 1410 inputs an image signal.
- This input image is, for example, image data obtained by a raster scan, as described above with reference to FIG.
- the image data includes, for example, a luminance signal in a component signal of a digital color video signal.
- Image data is input for each frame data, and is first stored in the current frame memory 1411 and then in the previous frame memory 1412. Therefore, the pixel correlation determination unit 1410 holds two consecutive frame data in the memory.
- pixel correlation determination section 1410 receives candidate vector information from candidate vector extraction section 102 shown in FIG. 6, and should determine each pixel of the previous frame, that is, a motion vector, based on the candidate vector information. For each pixel (pixel of interest), the pixel position of the current frame indicated by the plurality of candidate vectors is specified, and the difference between the pixel of interest of the previous frame and the pixel of the current frame indicated by the plurality of candidate vectors is calculated as the pixel value difference.
- the calculating unit 1413 calculates the difference, the absolute value calculating unit 1414 calculates the absolute value of the difference, and outputs the calculated absolute value to the temporary determining unit 1421.
- the tentative determination unit 1421 narrows down the motion vector corresponding to the pixel of interest from the n difference absolute values (MC residuals) [di]. Specifically, from the n ((difference absolute value (MC residual) [di]), the one with the small MC residual is selected. This processing is also the same as the above two-point matching. Assuming that the minimum value of the n MC residuals is dmin and the predetermined threshold is TH, di-dmm ⁇ TH
- the selection candidate level information is Only the candidate vector information selected from the candidate vector information output from the provisional determination unit 1421 to the gate 1433 and input from the candidate vector extraction unit 102 shown in FIG. 6 is input to the neighboring area information correlation determination unit 1430. Is done.
- the image correlation determining unit 1410 of the present embodiment further obtains the pixel information of the two pixels in the neighboring region of the pixel of interest from the current frame memory 1411 of the image correlation determining unit 1410 by the neighboring region information correlation determining unit 1430.
- the two pixels corresponding to the n designated pixels indicated by the candidate vector from the previous frame memory 1412 of the image correlation determining unit 1410, that is, n X 2 pixels, are input to the register 1432, that is, the neighboring area information correlation determination is performed. Input to register 1431 of unit 1430.
- Which pixel in the neighboring area is output is determined by the current frame additional information memory 1 422, and is determined by the preceding frame additional information memory 1423 and the output of n pieces of candidate vector information.
- the calculation result of the additional information calculation unit 1450 is output to the current frame additional information memory 1422 and the previous frame additional information memory 1423.
- the additional information calculation unit 1450 for example, among the pixels in the vicinity of the target pixel, the pixel having the largest pixel value difference from the target pixel, the pixel having the smallest pixel value difference from the target pixel, and these two pixels Extract as
- the attached information calculation unit 1450 shown in FIG. 36 includes a target pixel difference calculation unit 1451 including a register 1452 and a pixel value difference calculation unit 1453, a difference maximum value pixel information acquisition unit (MAX) 1454, and a difference minimum It has a value pixel information acquisition unit (MIN) 1455, a maximum difference pixel position detection unit 1456, a register 1457, a minimum difference pixel position detection unit 1458, and a register 1459.
- a target pixel difference calculation unit 1451 including a register 1452 and a pixel value difference calculation unit 1453
- a difference maximum value pixel information acquisition unit (MAX) 1454 a difference minimum It has a value pixel information acquisition unit (MIN) 1455, a maximum difference pixel position detection unit 1456, a register 1457, a minimum difference pixel position detection unit 1458, and a register 1459.
- MIN value pixel information acquisition unit
- the pixel-of-interest calculation unit 1451 inputs an image signal to be subjected to motion vector detection processing, and stores image frame data in the register 1452.
- the pixel value difference calculation unit 1453 sequentially calculates the difference between the neighboring areas of the image data stored in the register 1452, stores the pixel information having the maximum difference value in the maximum difference pixel information acquisition unit (MAX) 1454, The pixel information having the minimum difference value is stored in the minimum difference pixel information acquisition unit (MIN) 1455. Note that these pieces of pixel information include pixel position information.
- the additional information calculation unit 1450 sequentially calculates the difference between the pixel of interest and the pixel value near the pixel of interest with respect to the image data stored in the register 1452. And pixel information having a maximum difference value and pixel information having a minimum difference value.
- the pixels in the vicinity of the pixel of interest are, for example, a range of 1 8 1 + 7 pixels before and after the pixel of interest in the horizontal direction, or a block set in advance in a two-dimensional area. Pixel region.
- the pixel having the maximum difference value from the target pixel is the pixel 1482, that is, the pixel position [
- the pixel of [5] has the maximum difference value from the pixel of interest, that is, the maximum of the spatial gradient
- Pixel information is obtained as one selected characteristic pixel.
- the pixel having the minimum difference value from the target pixel is a pixel 1483 when the pixel position of the target pixel 1481 is [0], that is, the pixel having the minimum difference value from the target pixel, ie, the pixel at the pixel position [1], Pixel information is obtained as one feature pixel selected as the minimum pixel of the spatial gradient.
- the maximum difference pixel information obtaining unit (MAX) 1454 shown in Fig. 36 obtains a difference between a pixel of interest in a predetermined area, for example, every 16 pixels of -8 + 7, based on difference information of sequentially input pixels. The pixel having the maximum value is detected.
- a minimum difference pixel information acquisition unit (MIN) 1455 detects a pixel having a minimum difference value with respect to a target pixel for each predetermined area based on difference information of sequentially input pixels.
- the pixel information having the maximum pixel difference value for each predetermined area detected by the maximum difference pixel information acquisition unit (MAX) 1454 is input to the maximum difference pixel position detection unit 1456, and is output for each predetermined area (for example, 16 pixels).
- the maximum difference pixel position information is detected in each case and stored in the register 1457.
- the pixel information having the minimum pixel difference value for each predetermined area detected by the minimum difference pixel information acquisition unit (MIN) 1455 is input to the minimum difference pixel position detection unit 1458, and is output for each predetermined area (for example, 16 pixels). Is detected and stored in the register 1459.
- the pixel position of the pixel 1482 at which the difference from the target pixel is the largest is defined as:
- the pixel position of the pixel 1483 whose difference from the target pixel is the smallest is the pixel position [1] when the target pixel is set to the pixel position: 0.
- the maximum difference pixel position information for each predetermined region (for example, every 16 pixels) stored in the register 1457 and the minimum difference pixel position information for each predetermined region (for example, every 16 pixels) stored in the register 1459 are stored in the current frame. It is sequentially output to the additional information storage memory 1422.
- the characteristic pixel position information from the additional information calculation unit 1450 described with reference to FIG. 36 is input to the current frame additional information memory 1422 shown in FIG.
- the additional information moves to the previous frame additional information memory 1423 for each frame processing.
- Additional information of the current frame memory input to the current frame additional information memory 1422 that is, That is, two pieces of characteristic pixel position information corresponding to the target pixel, that is, the pixel position information having the maximum difference from the target pixel and the pixel position information having the minimum difference from the target pixel are stored in the current frame memory of the pixel correlation determination unit 1410.
- the information is output to 1411, and based on the output information, the two points of pixel information are stored in the register 1432 of the neighboring area information determination unit 1430.
- This pixel information includes pixel position information.
- the additional information of the previous frame memory input to the previous frame additional information memory 1423 that is, the pixel position information having the maximum difference from the target pixel and the pixel position information having the minimum difference from the target pixel are output to the previous frame memory 1412 of the pixel correlation determination unit 1410.
- the previous frame memory 1412 further receives n pieces of candidate vector information from the candidate vector extracting unit, and specifies two characteristic pixels, a neighboring area force determined for each pixel indicated by the candidate vector information, and n X
- the second characteristic pixel information is stored in the register 1431 of the neighboring area information determination unit 1430. This pixel information includes pixel position information.
- n X 2 characteristic pixel information stored in the register 1431 is output to the correlation determination unit 1434 via the gate 1433.
- the gate 1433 based on the selection candidate vectors selected by the provisional judgment processing in the provisional judgment unit 1421 from the n candidate vectors, only the characteristic pixel information in the vicinity area of the pixel corresponding to the selection candidate vector is correlated. Output to unit 1434.
- the characteristic pixel information of the target pixel is output to the direct correlation determination unit 1434.
- the correlation determination unit 1434 performs a correlation determination process on these characteristic pixels, and determines only one motion vector from the selection candidate vectors.
- FIG. 38 shows input frame (current frame) image data 1500 and past frame (previous frame) image data 1510 similar to those described above with reference to FIG.
- the input frame (current frame) image data 1500 shows candidate vectors a, c, and e selected by the candidate vector selection processing by the tentative determination unit described above with reference to FIG.
- the correlation determining unit 1433 in the neighboring area information determining unit 1430 performs the processing in the past frame (previous frame). Based on the pixel of interest selected from the image data 1510, that is, the characteristic pixel position data in the vicinity of the pixel of interest 1511 as the pixel to which the motion vector is associated, one motion vector such as the candidate vector a, c, e To determine.
- the area near the pixel of interest 1511 is a two-dimensional block area.
- the setting of this neighboring area is arbitrary, and may be set to a one-dimensional area in the horizontal direction, similarly to the above-described two-point matching.
- the spatial gradient maximum pixel 1512 which is the pixel position that is the maximum difference pixel from the pixel of interest
- the spatial gradient minimum pixel which is the pixel position that is the minimum difference pixel from the pixel of interest 1513
- the spatial gradient maximum pixel that is the pixel position that is the maximum differential pixel from the central pixel and the spatial gradient that is the pixel position that is the minimum differential pixel from the central pixel is determined.
- the correlation determination unit 1433 determines the correlation between the characteristic pixel position in the vicinity of the target pixel 1511 and the characteristic pixel position in the vicinity of the corresponding pixel position indicated by the selection candidate vectors a, c, and e.
- the neighborhood area 1501 indicated by the selection candidate vector a is located at a position corresponding to the characteristic pixel position of the neighborhood area of the target pixel 1511, and the other two selection candidate vectors c , e indicate that the neighboring areas 1502 and 1503 do not correspond to the characteristic pixel positions of the neighboring area of the target pixel 1511.
- a candidate vector a is selected and determined as a motion vector corresponding to the pixel of interest 1511.
- the correlation determination unit 1433 sequentially sets such processing as constituent pixels of a frame as pixels of interest, and performs correlation determination processing of characteristic pixel positions in the vicinity area in the same manner as described above for each set pixel of interest. Then, a motion vector corresponding to each pixel constituting the frame is selected and determined from the selection candidate vectors. [0337] As described above, the motion vector determination unit 400 shown in FIG. 35 specifies two characteristic pixel positions based on the pixel values in the region near the target pixel, and focuses on the characteristic pixels in the region near the target pixel.
- the correlation between the relative position corresponding to the pixel position and the characteristic pixel position in the vicinity of the pixel indicated by the candidate vector selected by the tentative judgment unit 1221 is determined, and the pixel having the characteristic pixel position with the highest correlation is specified. Is determined as the motion vector corresponding to the pixel of interest.
- the above-described motion vector determination processing sequence will be described with reference to the flowchart shown in FIG.
- the flow shown in FIG. 39 is a process of determining a motion vector for one pixel of interest.When determining motion vectors for all of the pixels constituting the frame image, the pixels constituting the frame image are sequentially set as pixels of interest. The flow shown in FIG. 39 is executed for each pixel of interest.
- step S1301 is a process of calculating the absolute value of the difference between the pixel of interest and the pixels specified by n candidate vectors, that is, the MC residual dn. This process is executed in the pixel correlation determination unit 1410 of the motion vector determination unit 400 in FIG.
- Steps S1302 and S1304 are processing in the provisional judgment section 1421.
- the tentative determination unit 1421 receives the n absolute values of the difference (MC residual) [di] and detects the minimum value of the n residuals: dmin.
- step S1304 is a tentative determination process executed as a process of selecting from candidate vectors described above with reference to Fig. 28. If it is determined that di? dmin ⁇ TH is not satisfied, it is determined that the vector is not a selection candidate vector, and the correlation determination based on the characteristic pixel position in the vicinity of the target pixel performed in steps S1305, S1306, and S1307 is performed. Otherwise, go to step S1308.
- step S1304 di? Dmin ⁇ TH is satisfied. In this case, the selection is made as a selection candidate vector, and the process proceeds to step S1305.
- step S1305 the correlation between characteristic pixel positions is determined. As described above with reference to FIG. 36, this processing is based on the correlation between the positions of two feature pixels in the vicinity of the pixel of interest and the two feature pixels in the vicinity of the pixel selected as the selection candidate vector. That is, this is a process of determining whether or not there is a correspondence between positions.
- step S1306 a process of selecting a higher correlation and a higher correlation is executed, and in step S1307, it is determined whether or not all candidate betatonets have been verified based on the value of i. If not, the process proceeds to step S1310, where the value of i is updated (incremented by one), and the processes in step S1304 and thereafter are repeatedly executed.
- step S1308 the two feature pixels in the vicinity of the target pixel calculated in step S1305 that have the highest correlation are selected. And a candidate vector corresponding to a pixel having a high correlation between the characteristic pixel position in the neighborhood of the pixel selected as the selection candidate vector and a motion vector corresponding to the pixel of interest to be processed.
- the correlation determination is performed based on the position information by applying the two characteristic pixel positions of the current frame and the previous frame, but the pixel value at the characteristic pixel position corresponding to the target pixel is determined. It may be configured to perform a correlation determination process based on the information. That is, block matching is performed using only a total of three points using the pixel value of the pixel of interest and the two characteristic pixels in the neighboring area. This processing sequence will be described with reference to FIG.
- FIG. 40 is a diagram showing input frame (current frame) image data 1600 as in FIG. 38 described above.
- the input frame (current frame) image data 1600 has been described with reference to FIG.
- the candidate vectors a, c, and e selected by the candidate vector selection processing by the tentative determination unit are shown.
- the correlation determination unit 1433 in the neighboring area information determination unit 1430 includes a feature in the neighborhood of the pixel of interest 1511 as a pixel selected from the past frame (previous frame) image data 1510, that is, a pixel to which a motion vector is associated. Based on the pixel position data, one candidate vector a, c, e force, etc., is determined as one motion vector.
- the area near the pixel of interest 1611 is a two-dimensional block area.
- the setting of this neighboring area is arbitrary, and may be set to a one-dimensional area in the horizontal direction, similarly to the above-described two-point matching.
- the spatial gradient maximum pixel 1612 which is the pixel position that is the maximum difference pixel from the pixel of interest
- the spatial gradient minimum pixel that is the pixel position that is the minimum difference pixel from the pixel of interest 1613
- a pixel region set around the pixel specified by each selection candidate vector that is, a pixel region 1601 set by the selection candidate vector a
- a selection candidate In each of the pixel region 1602 set by the vector c and the pixel region 1603 set by the selection candidate vector e the position of the maximum spatial gradient pixel 1612 and the minimum spatial gradient pixel 1613 with respect to the pixel of interest 1611 detected in the past frame are picked up and output to the neighboring area information correlation determination unit 1430 shown in FIG.
- the neighborhood value information correlation determination unit 1430 shown in FIG. 35 performs pixel value correlation determination between these pixels. That is, the correlation determination is performed between the pixel at each position of the spatial gradient maximum pixel 1612 and the spatial gradient minimum pixel 1613 with respect to the pixel of interest 1611 and the pixel values at the corresponding positions of the pixel regions 1601, 1602, and 1603. For the center pixel, correlation determination was performed in the provisional determination unit 1421, and as a result, only three points were used. The matching processing result is obtained, and the selection candidate vector corresponding to the data determined to have the highest correlation by the three-point matching processing is determined as the motion level corresponding to the pixel of interest.
- Fig. 41 shows a flow of the motion vector determination processing according to the above method.
- the processing of steps S1401 and S1404 is the same as the processing of steps S1301 to S1304 of the processing flow described above with reference to FIG. 39, and the selection candidate vector in the temporary determination unit is Is the process to be extracted.
- step S1405 block matching using three pixels is executed. That is, as described with reference to FIG. 40, from each pixel area indicated by the selection candidate vector, the position corresponding to the position of the spatial gradient maximum pixel and the position of the spatial gradient minimum pixel for the pixel of interest detected in the past frame is determined. The pixel to be processed is picked up and the correlation is determined. For the central pixel, that is, the pixel correlation information corresponding to the target pixel position, the determination result of the temporary determination unit is referred to.
- Steps S1406—S1408 and S1410 are the same as steps S1306—S1308 and S1310 of the processing flow described above with reference to FIG.
- the target pixel without performing the correlation determination using all the pixels in the vicinity (block) of the target pixel is determined.
- two characteristic pixel position information and pixel value information in the vicinity are used. Therefore, it is not necessary to perform a correlation calculation based on a large number of pixel values according to the size of a block to be applied in the block matching processing, and efficient processing can be performed.
- Fig. 42 illustrates a processing sequence in which the correlation determination using the correlation information of the characteristic pixel position and the motion level determination processing based on the correlation determination using the correlation information of the pixel values are used together. Indicates low.
- the processing in steps S1501 to S1504 is the same as the processing in steps S1301 to S1304 in the processing flow described above with reference to Fig. 39, and the selection candidate vector in the tentative determination unit Is the process to be extracted.
- step S1505 the correlation between the characteristic pixel positions is determined. As described above with reference to FIG. 36, this processing is based on the correlation between the positions of two feature pixels in the vicinity of the pixel of interest and the two feature pixels in the vicinity of the pixel selected as the selection candidate vector. That is, this is a process of determining whether or not there is a correspondence between positions.
- step S1506 block matching using three pixels is performed. That is, as described with reference to FIG. 40, from each pixel region indicated by the selection candidate vector, the position of the spatial gradient maximum pixel relative to the target pixel detected in the past frame and the position of the spatial gradient minimum pixel are determined. The pixel corresponding to is picked up and the correlation is determined. For the pixel correlation information corresponding to the center pixel, that is, the pixel position of interest, the determination result of the temporary determination unit is referred to.
- step S1507 the vector indicating the highest level and the correlation is selected in consideration of both the correlation half-IJ determination in step S1505 and the correlation determination in step S1506.
- the processing in steps S1508 to S1510 is the same as the processing in steps S1307 to S1308 and S1310 of the processing flow described above with reference to FIG. 39, and a description thereof will be omitted.
- the program is stored on a hard disk or ROM (Read Only Memory
- the program can be temporarily stored on a removable recording medium such as a flexible disk, CD-R ⁇ M (Compact Disc Read Only Memory), MO (Magneto optical) disk, DVD (Digital Versatile Disc), magnetic disk, or semiconductor memory. Alternatively, it can be permanently stored (recorded).
- a removable recording medium can be provided as a package software.
- the program can be wirelessly transferred from a download site to the computer, or transmitted to the computer via a network such as a LAN (Local Area Network) or the Internet.
- the program can be transferred by wire, and the computer can receive the transferred program and install it on a recording medium such as a built-in hard disk.
- the correlation considering not only the representative point but also the spatial waveform as the pixel level near the representative point is considered. Since the determination process is performed, it is possible to generate an evaluation value table with higher accuracy, and it is possible to more accurately execute motion vector detection.
- the correlation information of the representative point pixel, the target pixel and the target pixel A weighting factor W is calculated based on flag correlation information based on flag data corresponding to pixel value difference data of neighboring pixels, and the calculated weighting factor W and activity A as an index value indicating the complexity of image data are calculated.
- a reliability index is calculated as a calculated value based on, and an evaluation value table in which evaluation values corresponding to the reliability index are integrated is generated.
- the configuration is such that the flag data corresponding to the pixel value difference data of the target pixel and the pixel in the vicinity of the target pixel is calculated as a bit number smaller than the pixel value difference data.
- the calculation of the flag correlation calculation process can be performed as a process with a small number of bits, and a highly accurate evaluation value table can be generated without reducing the processing speed, and the motion vector detection can be performed more accurately. It is possible to do. Therefore, by applying the present invention to an image processing device or the like that performs a coding process of moving image data, an efficient motion vector detection process can be performed, and the device can be downsized.
- the motion similarity between the representative point and the pixel near the representative point is determined, and when there is motion similarity.
- a value table can be generated, and motion vector detection can be performed more accurately.
- the motion similarity between the representative point and the representative point in the vicinity of the representative point is detected, and the representative point and the neighboring representative point are determined on condition that the motion similarity is determined.
- This is a configuration that generates a reliability index / 3 taking into account the pixel value difference from the point, and generates an evaluation value table in which the evaluation values corresponding to the reliability index ⁇ are integrated.
- This configuration calculates the reliability index j3 in consideration of these parameters, and generates the evaluation value table by integrating the evaluation values based on the reliability index.Thus, it is possible to generate a more accurate evaluation value table. Thus, it is possible to more accurately execute motion vector detection.
- the flag correlation information based on the flag data corresponding to the pixel value difference data of the target pixel and the pixel in the vicinity of the target pixel is further added.
- a weighting factor W is calculated based on the calculated weighting factor W
- An evaluation value table based on the evaluation values can be generated, and a more accurate evaluation value table can be generated, and motion vector detection can be performed more accurately. Therefore, by applying the present invention to an image processing device or the like that executes a coding process of moving image data and the like, an efficient motion vector detection process can be performed, and the device can be downsized.
- a process of selecting and determining a motion vector corresponding to each pixel from a plurality of candidate vectors is used.
- a feature pixel is extracted from a region near the target pixel to be associated, a correlation determination is performed based on the position information or the pixel value information of the feature pixel, and a process of determining a motion vector is performed. It is not necessary to apply it.As calculation for correlation calculation, it is possible to reduce both evaluation value calculation such as summation of absolute difference values, thereby improving processing efficiency and reducing memory for holding pixel values. And the hardware size can be reduced. Therefore, by applying the present invention to an image processing device or the like that executes encoding processing of moving image data, more accurate motion vector detection processing becomes possible.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
Claims
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020067016529A KR101098394B1 (ko) | 2004-03-01 | 2006-08-17 | 움직임 벡터 검출 장치 및 움직임 벡터 검출 방법 |
Applications Claiming Priority (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004056255A JP4626158B2 (ja) | 2004-03-01 | 2004-03-01 | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム |
JP2004-056255 | 2004-03-01 | ||
JP2004-056257 | 2004-03-01 | ||
JP2004056256A JP4622264B2 (ja) | 2004-03-01 | 2004-03-01 | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム |
JP2004-056256 | 2004-03-01 | ||
JP2004056257A JP4622265B2 (ja) | 2004-03-01 | 2004-03-01 | 動きベクトル検出装置、および動きベクトル検出方法、並びにプログラム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2005084036A1 true WO2005084036A1 (ja) | 2005-09-09 |
Family
ID=34916091
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2005/002786 WO2005084036A1 (ja) | 2004-03-01 | 2005-02-22 | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US8064522B2 (ja) |
KR (1) | KR101098394B1 (ja) |
WO (1) | WO2005084036A1 (ja) |
Families Citing this family (56)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1921867B1 (en) * | 2006-10-17 | 2016-05-25 | Harman Becker Automotive Systems GmbH | Sensor assisted video compression |
US8159605B2 (en) * | 2007-07-13 | 2012-04-17 | Fujitsu Limited | Frame interpolating apparatus and method |
WO2009032255A2 (en) * | 2007-09-04 | 2009-03-12 | The Regents Of The University Of California | Hierarchical motion vector processing method, software and devices |
JP5045320B2 (ja) * | 2007-09-05 | 2012-10-10 | ソニー株式会社 | 画像処理装置、および画像処理方法、並びにコンピュータ・プログラム |
US20090141802A1 (en) * | 2007-11-29 | 2009-06-04 | Sony Corporation | Motion vector detecting apparatus, motion vector detecting method, and program |
US8743952B2 (en) * | 2007-12-18 | 2014-06-03 | Vixs Systems, Inc | Direct mode module with motion flag precoding and methods for use therewith |
KR101540138B1 (ko) * | 2007-12-20 | 2015-07-28 | 퀄컴 인코포레이티드 | 적응적 조사영역을 갖는 모션추정 장치 및 방법 |
KR101536794B1 (ko) * | 2007-12-20 | 2015-07-14 | 퀄컴 인코포레이티드 | 후광현상이 줄어든 영상보간 장치 및 방법 |
US8351509B1 (en) * | 2007-12-21 | 2013-01-08 | Dimension, Inc. | Video codec systems and methods for optimal motion vector with variable size blocks |
WO2009096169A1 (ja) * | 2008-01-29 | 2009-08-06 | Panasonic Corporation | モータ制御特性評価装置およびモータ制御特性の提示方法 |
US8331696B2 (en) * | 2008-02-14 | 2012-12-11 | Nec Corporation | Update region detection device |
JP5333440B2 (ja) * | 2008-03-28 | 2013-11-06 | 日本電気株式会社 | 処理システム、画像処理方法および画像処理用プログラム |
JP2010011016A (ja) * | 2008-06-26 | 2010-01-14 | Sony Corp | 追尾点検出装置および方法、プログラム、並びに記録媒体 |
KR101364195B1 (ko) | 2008-06-26 | 2014-02-21 | 에스케이텔레콤 주식회사 | 움직임벡터 부호화/복호화 방법 및 그 장치 |
TWI470578B (zh) * | 2008-06-26 | 2015-01-21 | Mstar Semiconductor Inc | 計算用於待插補影像區塊之目標移動向量的影像處理方法及其相關裝置 |
JP4697276B2 (ja) * | 2008-07-30 | 2011-06-08 | ソニー株式会社 | 動きベクトル検出装置、動きベクトル検出方法及びプログラム |
JP4748191B2 (ja) * | 2008-07-30 | 2011-08-17 | ソニー株式会社 | 動きベクトル検出装置、動きベクトル検出方法及びプログラム |
JP4697275B2 (ja) * | 2008-07-30 | 2011-06-08 | ソニー株式会社 | 動きベクトル検出装置、動きベクトル検出方法及びプログラム |
JP5245783B2 (ja) * | 2008-12-09 | 2013-07-24 | 富士通株式会社 | フレーム補間装置、方法、及びプログラム、フレームレート変換装置、映像再生装置、映像表示装置 |
JP2010114596A (ja) * | 2008-11-05 | 2010-05-20 | Sony Corp | 動きベクトル検出装置、その処理方法およびプログラム |
CN101739400B (zh) * | 2008-11-11 | 2014-08-13 | 日电(中国)有限公司 | 生成索引的方法和装置以及检索方法和装置 |
US8509552B1 (en) * | 2009-03-09 | 2013-08-13 | Marvell International Ltd. | Digital image processing error concealment method |
JP2010288098A (ja) * | 2009-06-12 | 2010-12-24 | Sony Corp | 画像フレーム補間装置、画像フレーム補間方法及び画像フレーム補間プログラム |
JP5538967B2 (ja) | 2009-06-18 | 2014-07-02 | キヤノン株式会社 | 情報処理装置、情報処理方法、プログラム |
BR112012000002B1 (pt) * | 2009-07-03 | 2021-11-23 | France Telecom | Método para predizer um vetor de movimento de uma partição de imagem corrente, método para codificar uma imagem, método para decodificar um fluxo de dados representativo de uma imagem, dispositivo para predizer um vetor de movimento de uma partição de imagem corrente, dispositivo para codificar uma imagem e dispositivo para decodificar um fluxo de dados |
US8537283B2 (en) | 2010-04-15 | 2013-09-17 | Qualcomm Incorporated | High definition frame rate conversion |
US8670073B2 (en) * | 2010-08-02 | 2014-03-11 | Broadcom Corporation | Method and system for video noise filtering |
JP5639444B2 (ja) * | 2010-11-08 | 2014-12-10 | キヤノン株式会社 | 動きベクトル生成装置、動きベクトル生成方法及びコンピュータプログラム |
US9516389B1 (en) * | 2010-12-13 | 2016-12-06 | Pixelworks, Inc. | On screen display detection |
TWI444049B (zh) * | 2011-01-18 | 2014-07-01 | Univ Nat Chiao Tung | 畫面估測系統及其估測方法 |
JP2012151796A (ja) * | 2011-01-21 | 2012-08-09 | Sony Corp | 画像処理装置と画像処理方法およびプログラム |
US9532050B2 (en) * | 2011-01-25 | 2016-12-27 | Sun Patent Trust | Moving picture coding method and moving picture decoding method |
JP5887770B2 (ja) * | 2011-09-05 | 2016-03-16 | 富士ゼロックス株式会社 | 画像処理装置及び画像処理プログラム |
KR101544390B1 (ko) | 2013-06-26 | 2015-08-17 | 에스케이텔레콤 주식회사 | 움직임벡터 부호화/복호화 방법 및 그 장치 |
KR101567963B1 (ko) * | 2014-01-03 | 2015-11-11 | 에스케이텔레콤 주식회사 | 움직임벡터 부호화/복호화 방법 및 그 장치 |
KR101567990B1 (ko) * | 2014-01-03 | 2015-11-11 | 에스케이텔레콤 주식회사 | 움직임벡터 부호화/복호화 방법 및 그 장치 |
KR101567964B1 (ko) * | 2014-01-03 | 2015-11-11 | 에스케이텔레콤 주식회사 | 움직임벡터 부호화/복호화 방법 및 그 장치 |
CN104899853B (zh) | 2014-03-04 | 2019-12-06 | 腾讯科技(深圳)有限公司 | 图像区域的划分方法及装置 |
JP6223899B2 (ja) * | 2014-04-24 | 2017-11-01 | 株式会社東芝 | 動きベクトル検出装置、距離検出装置および動きベクトル検出方法 |
US10140827B2 (en) | 2014-07-07 | 2018-11-27 | Google Llc | Method and system for processing motion event notifications |
US9170707B1 (en) | 2014-09-30 | 2015-10-27 | Google Inc. | Method and system for generating a smart time-lapse video clip |
US10127783B2 (en) | 2014-07-07 | 2018-11-13 | Google Llc | Method and device for processing motion events |
US9213903B1 (en) | 2014-07-07 | 2015-12-15 | Google Inc. | Method and system for cluster-based video monitoring and event categorization |
US9449229B1 (en) | 2014-07-07 | 2016-09-20 | Google Inc. | Systems and methods for categorizing motion event candidates |
US9501915B1 (en) * | 2014-07-07 | 2016-11-22 | Google Inc. | Systems and methods for analyzing a video stream |
USD782495S1 (en) | 2014-10-07 | 2017-03-28 | Google Inc. | Display screen or portion thereof with graphical user interface |
US9361011B1 (en) | 2015-06-14 | 2016-06-07 | Google Inc. | Methods and systems for presenting multiple live video feeds in a user interface |
CN115278228A (zh) | 2015-11-11 | 2022-11-01 | 三星电子株式会社 | 对视频进行解码的方法和对视频进行编码的方法 |
US10506237B1 (en) | 2016-05-27 | 2019-12-10 | Google Llc | Methods and devices for dynamic adaptation of encoding bitrate for video streaming |
US10380429B2 (en) | 2016-07-11 | 2019-08-13 | Google Llc | Methods and systems for person detection in a video feed |
US11783010B2 (en) | 2017-05-30 | 2023-10-10 | Google Llc | Systems and methods of person recognition in video streams |
US10664688B2 (en) | 2017-09-20 | 2020-05-26 | Google Llc | Systems and methods of detecting and responding to a visitor to a smart home environment |
JP7009252B2 (ja) * | 2018-02-20 | 2022-01-25 | キヤノン株式会社 | 画像処理装置、画像処理方法およびプログラム |
KR102564334B1 (ko) * | 2018-08-03 | 2023-08-08 | 엘지디스플레이 주식회사 | 영상 처리 장치 및 방법, 디스플레이 장치 |
US11523109B2 (en) * | 2019-05-02 | 2022-12-06 | Tencent America LLC | Method and apparatus for improvements of affine prof |
US11188756B2 (en) * | 2019-10-16 | 2021-11-30 | Realtek Singapore Private Limited | Object localization and classification system and method thereof |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07118784B2 (ja) * | 1984-10-27 | 1995-12-18 | ソニー株式会社 | テレビジヨン信号の動き検出方法 |
JP2001061152A (ja) * | 1999-08-23 | 2001-03-06 | Sony Corp | 動き検出方法および動き検出装置 |
JP2002290977A (ja) * | 2001-03-26 | 2002-10-04 | Mitsubishi Electric Corp | 動きベクトル検出装置 |
JP2003078807A (ja) * | 2001-08-31 | 2003-03-14 | Sony Corp | 動きベクトル検出装置および方法、手振れ補正装置および方法、並びに撮像装置 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2924609B2 (ja) | 1993-10-26 | 1999-07-26 | 日本軽金属株式会社 | 鋼構造物防食用アルミニウム合金 |
KR0181031B1 (ko) | 1995-03-20 | 1999-05-01 | 배순훈 | 움직임 보상된 인터폴레이션에서의 엣지 보상 장치 |
US6532264B1 (en) * | 2000-03-27 | 2003-03-11 | Teranex, Inc. | Processing sequential video images to detect image motion among interlaced video fields or progressive video images |
-
2005
- 2005-02-22 WO PCT/JP2005/002786 patent/WO2005084036A1/ja active Application Filing
-
2006
- 2006-08-17 KR KR1020067016529A patent/KR101098394B1/ko not_active IP Right Cessation
- 2006-08-28 US US11/467,777 patent/US8064522B2/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07118784B2 (ja) * | 1984-10-27 | 1995-12-18 | ソニー株式会社 | テレビジヨン信号の動き検出方法 |
JP2001061152A (ja) * | 1999-08-23 | 2001-03-06 | Sony Corp | 動き検出方法および動き検出装置 |
JP2002290977A (ja) * | 2001-03-26 | 2002-10-04 | Mitsubishi Electric Corp | 動きベクトル検出装置 |
JP2003078807A (ja) * | 2001-08-31 | 2003-03-14 | Sony Corp | 動きベクトル検出装置および方法、手振れ補正装置および方法、並びに撮像装置 |
Also Published As
Publication number | Publication date |
---|---|
US8064522B2 (en) | 2011-11-22 |
KR20060132911A (ko) | 2006-12-22 |
US20060285596A1 (en) | 2006-12-21 |
KR101098394B1 (ko) | 2011-12-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2005084036A1 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム | |
KR101141467B1 (ko) | 화상 처리 장치 및 방법, 및 기록 매체 | |
US8285045B2 (en) | Image analysis method, medium and apparatus and moving image segmentation system | |
US7050502B2 (en) | Method and apparatus for motion vector detection and medium storing method program directed to the same | |
KR101106419B1 (ko) | 화상 처리 장치 및 방법, 및 기록 매체 | |
JP5338684B2 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
JP5577793B2 (ja) | 画像処理装置および方法、並びにプログラム | |
JP5163429B2 (ja) | 動きベクトル検出装置、その処理方法およびプログラム | |
US20050180506A1 (en) | Unit for and method of estimating a current motion vector | |
JP2010034997A (ja) | 動きベクトル検出装置、動きベクトル検出方法及びプログラム | |
JP5429564B2 (ja) | 画像処理装置および方法、並びにプログラム | |
JP4626158B2 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム | |
JP2011146827A (ja) | 画像処理装置および方法、並びにプログラム | |
JP2009533887A (ja) | 動きベクトル場修正装置及び方法 | |
JP2009212605A (ja) | 情報処理方法、情報処理装置及びプログラム | |
JP2011146826A (ja) | 画像処理装置および方法、並びにプログラム | |
JP2011211629A (ja) | 画像処理装置および方法、並びにプログラム | |
JP4525064B2 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム | |
JP4622264B2 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム | |
JP4622265B2 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにプログラム | |
US9106926B1 (en) | Using double confirmation of motion vectors to determine occluded regions in images | |
JP3175914B2 (ja) | 画像符号化方法および画像符号化装置 | |
US10063880B2 (en) | Motion detecting apparatus, motion detecting method and program | |
JP4207764B2 (ja) | 動きベクトル検出装置、および動きベクトル検出方法、並びにコンピュータ・プログラム | |
KR101403143B1 (ko) | 컨텐츠에 대한 관객 그룹의 반응 판단 방법 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 1020067016529 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200580006359.9 Country of ref document: CN |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
WWP | Wipo information: published in national office |
Ref document number: 1020067016529 Country of ref document: KR |
|
122 | Ep: pct application non-entry in european phase |