JP2002223419A - Sequential scanning conversion method and sequential scanning converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sequential scanning conversion method and a sequential scanning converter that can enhance the performance of a VT filter applied to a moving picture by the sequential scanning conversion. SOLUTION: A VT filter 102 receives data of a field #n that is a field of an object of sequential scanning conversion and data of fields #n-1 and #n+1 that are fields before and after the field #n. A differential arithmetic unit 104 receives data by two frames including the field #n and calculates a differential absolute sum of the frames. A filter coefficient setting device 103 decides a filter coefficient on the basis of the differential absolute sum. The VT filter 102 uses the filter coefficient to apply filter processing to an input pixel and to generated and output an interpolation pixel. A double speed converter 105 synthesizes an interlace image with the interpolation pixel and revises the image to have a double frame rate and outputs the result as a sequential image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a progressive scan conversion method and a progressive scan conversion apparatus for converting an interlaced image signal into a progressive image signal.

[0002]

2. Description of the Related Art In recent years, digital processing of images has become widespread. As a method of compressing and encoding image data,
High-efficiency coding such as MPEG (Moving Picture Experts Group) is adopted. The high-efficiency coding technique is a technique for coding image data at a low bit rate in order to improve the efficiency of digital transmission and recording.

[0003] In the current NTSC television broadcasting, an interlaced scanning system is adopted. Interlaced scanning is one screen (one frame)
To the top (top) field and the bottom (botto)
m) By dividing the data into fields and transmitting the data, the bandwidth can be saved and the transmission can be performed with high efficiency. However, in the interlaced scanning method, line flicker and line crawl are conspicuous due to the high brightness and large screen of the television receiver. Therefore, by performing interpolation using an image memory, the interlaced image signal is converted into a progressive image signal. May be performed.

[0004] An example of a progressive scan conversion method for converting an interlaced image signal into a progressive image signal is disclosed in Japanese Patent Application Laid-Open No. 7-131761.
This method is also called a VT filter method. In the VT filter method, an interpolated pixel is generated by performing filter processing on a pixel in a field in which an interpolated pixel is generated and pixels in three consecutive fields including a field before and after the interpolated pixel, and using the generated interpolated pixel. Convert an interlaced image to a progressive image. Hereinafter, this VT filter method will be described with reference to FIG.

FIG. 25 schematically shows the state of the time-vertical plane of an interlaced image, in which white circles indicate pixels (scanning lines). Here, a description will be given of the filter processing in the case where the pixel at the position k ′ shown in FIG. 25 is generated by interpolation.

In this case, in the VT filter system, pixels in a field including k 'and pixels in three consecutive fields before and after the field k, for example, 10 pixels a' to j 'as shown in FIG. On the other hand, the pixel at the position k ′ is generated by performing the filtering process. At this time, as a filter used for the filter processing, pixels (pixels d ′ to g ′) in the same field have a low-pass characteristic in the vertical direction, and pixels (pixels a ′ to
For c ′, h ′ to j ′), a filter having a vertical high-pass characteristic is used. A filter having such characteristics extracts a vertical low-pass component from pixels d ′ to g ′, and a vertical high-pass component from pixels a ′ to c ′ and h ′ to j ′. I do. In the VT filter method, an interpolation pixel is generated by adding the low-pass component and the high-pass component extracted by the filter.

As another example of the progressive scan conversion method, there is a method disclosed in Japanese Patent Application Laid-Open No. H11-331782 (reference 2). The method described in this publication includes a still image filter, a moving image filter, a motion detection circuit,
This is a method realized by a signal conversion device having a mixer. In this method, first, the still image filter and the moving image filter each generate an interpolation pixel. Next, based on the detection result of the motion detector that detects the motion of the image, the mixer generates the final interpolated pixel by mixing the interpolation signals output from the still image filter and the moving image filter. Then, the interlaced image is converted into a progressive image by the final interpolated pixels. The motion detection circuit obtains a difference between frames, and obtains a motion amount from an absolute value of the difference.

Hereinafter, this progressive scan conversion method will be described with reference to FIG.
5 will be described. Here, a description will be given of the filter processing in the case where the pixel at the position k ′ shown in FIG. 25 is generated by interpolation.

In this method, first, a still image filter performs filter processing on the pixels b ′, e ′, f ′, and i ′ to generate an interpolated pixel k ′, and the moving image filter generates pixels a ′ to
j ′ is filtered to generate an interpolated pixel k ′. Note that the filter coefficients of the still image filter have the heaviest weights for the pixels b ', i', and the filter coefficients of the moving image filters have the heavier weights for the pixels d ', e', f ', g'. It is set to be. Next, the motion detector detects the motion of the image from the difference between the frames, and based on the detection result, the mixer outputs the interpolation signal output by the still image filter and the interpolation signal output by the moving image filter. Are mixed to generate a final interpolated pixel.

In the above description, the filtering process is performed on the pixels of three fields. However, the filtering process may be performed on the pixels of two fields.
In this case, the still image filter performs filter processing on e ′, f ′, and i ′, and the moving image filter performs filter processing on pixels d ′ to j ′ to generate an interpolation pixel k ′.

As another example of the progressive scan conversion method, a motion vector is detected between a field to be progressively scanned and its neighboring field, and interpolation is performed from neighboring fields according to the result of the motion vector detection. There is a method of acquiring pixels and converting an interlaced image to a progressive image using the interpolated pixels. Although this method detects the motion of an image as motion vector information, it enables interpolation processing that almost matches the motion of the image.However, detection of a motion vector requires a huge amount of signal processing. Was.

Therefore, in a method of performing progressive scan conversion using a motion vector, a progressive scan conversion apparatus which realizes a method of reducing the processing amount of motion vector detection is disclosed in Japanese Patent Laid-Open No. 10-1.
No. 26749 (Reference 3).

[0013] The progressive scan conversion device described in this publication has a M
An interlaced image signal (MPEG video) encoded by an encoding method such as PEG is input, and the MPEG
Circuits such as a motion detection circuit and a motion vector detection circuit are omitted by adaptively switching an interpolation method in the progressive scan conversion using information such as an image structure and a motion vector included in a video. Here, the MPEG video input to the progressive scan conversion device is an MPE
A case where the data is encoded by the G2 encoding method will be described.

The MPEG2 encoding process includes an intra-frame encoding process and an inter-frame encoding process. When intra-frame encoding is performed on an interlaced image,
An interlaced image is, for example, a block of 8 × 8 pixels (DC
After DCT (Discrete Cosine Transform) processing is performed in units of T blocks, and the spatial coordinate components are converted into frequency components, variable-length coding is performed.

On the other hand, when an inter-frame encoding process is performed on an interlaced image, the difference between the image of the current frame and the reference images of the preceding and succeeding frames is obtained as a prediction error. And variable-length coding. Thereby, the code amount can be significantly reduced. However, when the motion of the image is large, simply calculating the difference between the previous and next images may increase the prediction error and increase the code amount. Therefore, by detecting a motion vector between the reference image and the image of the current frame, performing motion compensation on the reference image based on the motion vector and obtaining a difference between the reference image and the image of the current frame, the prediction error is reduced and the code amount is reduced. Reduce. Note that M
In the PEG2 encoding process, one image (picture) consisting of one frame or one field is divided into macroblocks consisting of 16 lines × 16 pixels, a motion vector is detected in units of the macroblock, and motion compensation is performed. I have.

Among the images encoded by the above-described encoding processing, an intra-frame encoded image in which image data is predictively encoded in one frame is an I-picture, a predictive encoding using a forward frame. The inter-frame forward coded image in which the image data has been encoded is a P-picture, a forward frame, a backward frame, and both of which the image data has been encoded by predictive encoding using any of the bidirectional frames. A directional encoded image is called a B picture.

In the coding of the interlaced image, the coding of the top and bottom fields is performed as it is, that is, the coding is performed in the state of the field structure, and the coding of the top and bottom fields as the field image is performed. There is a case where a frame image is created from an image and encoding is performed in a state of a frame structure. In the coding in the state of the field structure, motion compensation prediction and DCT coding are performed for each field picture. On the other hand, encoding in the state of a frame structure is performed in units of frame pictures. As motion compensation prediction, a frame prediction in which a frame image is used as a reference image, and a top field image and Field prediction using one of the bottom field images is employed.

Accordingly, by decoding the MPEG video code string coded as described above, the progressive scan conversion apparatus can control the image structure, the motion vector, the prediction mode, etc., at the time of coding the interlaced image. Information can be obtained. Then, based on the obtained information, it is determined whether the decoded image data is a still area or a moving area. If it is determined that the decoded image data is a still area, inter-field interpolation is performed to determine that the image data is a moving area. In this case, the interlaced image signal is converted into a progressive image signal by performing intra-field interpolation.

However, as described above, in the progressive scan conversion, the vertical resolution of a progressive image can be improved by using inter-field interpolation rather than by using intra-field interpolation. Therefore, the above-described progressive scan conversion device determines whether inter-field interpolation is possible even when the decoded image data is a moving area. Here, the amount of motion of the motion vector at the time of encoding included in the MPEG video is obtained, and if the amount of motion is an even number of pixels, it is determined that inter-field interpolation is possible. Then, when it is determined that the inter-field interpolation is also possible in the moving area, the vertical resolution of the converted progressive image can be improved by performing the inter-field interpolation based on the motion vector.

[0020]

However, in the VT filter system described in the above-mentioned reference 1, an interpolated pixel is generated regardless of whether an interlaced image is a still image or a moving image. That is, the filtering process is performed on the pixels in the field where the interpolation pixel is always generated and the pixels in the three consecutive fields including the preceding and succeeding fields. Therefore, in the case of a still image, a progressive image having a high resolution can be generated. However, in the case of a moving image, since the correlation between a field for generating an interpolation pixel and a field before and after the field is low, filtering processing is performed on pixels in the previous and next fields. To generate an interpolation pixel, image quality may deteriorate. In particular, when an image including a diagonal edge moves vertically, deterioration occurs such that the edge appears to rattle.

Further, the method described in Document 2 requires two types of filters, a still image filter and a moving image filter, in order to generate an interpolated pixel. there were.

In general, in the MPEG2 encoding method, when an interlaced image is motion-compensated for predictive encoding, a motion vector between frames 1 to 3 frames apart is detected, and a reference image is calculated using the motion vector. Is motion compensated. Therefore, when the progressive scan converter performs the inter-field interpolation directly using the motion vector obtained at the time of decoding the MPEG video, the interpolated pixel may be obtained from a frame separated by a plurality of frames. . The motion vector used at the time of motion compensation for coding is
Even if the detection accuracy is poor, the coding efficiency will be slightly deteriorated, and the image quality will not be significantly degraded.Therefore, there is no problem even if the detection is performed between frames separated by a plurality of frames. In the progressive scan conversion that performs the interpolation of, it is necessary to detect a motion vector with higher accuracy. Therefore, in the progressive scanning change device described in the above-mentioned Document 3, if an interpolation pixel is obtained by directly using a motion vector at the time of motion compensation for encoding, an interpolation pixel is obtained from a frame separated by a plurality of frames. In some cases, interpolation pixels could not be generated.

The present invention solves the above-mentioned problems. With a small-scale circuit configuration, a high-resolution progressive image can be generated for a still image, and the image quality does not deteriorate even in a moving image. An object of the present invention is to provide a progressive scan conversion method and a progressive scan conversion device that can generate a progressive image.

Further, even in the case of performing the progressive scan conversion process using a motion vector obtained by decoding an MPEG video code sequence, the progressive scan conversion can generate an interpolation pixel with high accuracy. It is an object to provide a method and a progressive scan conversion device.

[0025]

According to a first aspect of the present invention, there is provided a progressive scan conversion method for converting an interlaced scan image into a progressive scan image. The filter processing is performed on pixels of at least one of three fields of a field to be subjected to the sequential scan conversion processing and a field before and after the field to be subjected to the sequential scan conversion, thereby performing the sequential scan conversion. Generating an interpolation pixel of the scan conversion target field, measuring a motion amount of the progressive scan conversion target field, and changing a characteristic of the filtering process based on the motion amount. And

In the progressive scan conversion method according to a second aspect of the present invention, in the progressive scan conversion method according to the first aspect, the filter used for the filtering in the step of generating the interpolated pixel is the same as the progressive scan conversion method. For a field to be subjected to scan conversion, a low frequency component in the vertical direction is extracted, and for fields before and after the field to be subjected to progressive scan conversion, the high frequency component in the vertical direction is extracted. Features.

According to a third aspect of the present invention, there is provided a progressive scan conversion method according to the first aspect, wherein the step of generating the interpolated pixel includes the step of: , Filtering is performed on peripheral pixels in the vertical-time plane.

According to a fourth aspect of the present invention, in the progressive scan conversion method of the first aspect, the step of measuring the amount of motion is performed in the field to be subjected to the progressive scan conversion or the progressive scan conversion. The motion amount is obtained from a difference value between a frame including the scan conversion target field and another field or frame.

In the progressive scan conversion method according to a fifth aspect of the present invention, in the progressive scan conversion method according to the first aspect, the step of measuring the amount of motion includes the step of generating the interpolation pixel. The amount of motion is obtained from a difference value between pixels used when performing a filtering process.

In the progressive scan conversion method according to a sixth aspect of the present invention, in the progressive scan conversion method according to the fifth aspect, the step of measuring the amount of motion includes the step of generating the interpolation pixel. A motion amount is obtained from a difference value between pixels belonging to fields before and after the field to be subjected to the progressive scan conversion among pixels used when applying a filter.

According to a seventh aspect of the present invention, there is provided a progressive scan conversion method according to the first aspect, wherein the step of changing the characteristics of the filter processing is performed by the following steps.
The characteristic of the filter processing is changed such that the gain of the component from the field before and after the field to be sequentially converted becomes smaller as the amount of motion increases.

In the progressive scan conversion method according to the present invention, in the progressive scan conversion method according to the first aspect, the step of changing the characteristic of the filter processing may include:
When the motion amount is large, the characteristic of the filter processing is changed so that the gain of the components from the fields before and after the field to be subjected to the progressive scan conversion becomes zero.

According to a ninth aspect of the present invention, there is provided a progressive scan conversion device for converting an interlaced scan image into a progressive scan image, comprising: a frame memory for accumulating the interlaced scan image; In the interlaced scan image, a progressive scan conversion target field to be subjected to progressive scan conversion processing, and one or both of a preceding field and a subsequent field of the progressive scan conversion field are input, and the input field is input. A filter for generating an interpolated pixel of the progressive scan conversion target field by applying a filter process to pixels of at least one of the fields; a difference calculator for measuring a motion amount of the progressive scan conversion target field; Changes the characteristics of the filter based on the amount of motion measured by the difference calculator A filter coefficient setting unit that, characterized in that it comprises a.

According to a tenth aspect of the present invention, there is provided a progressive scanning conversion device for converting an interlaced scanning image into a progressive scanning image, comprising: a frame memory for accumulating an interlaced scanning image; In the interlaced scan image, a progressive scan conversion target field to be subjected to progressive scan conversion processing, and one or both of a preceding field and a subsequent field of the progressive scan conversion target field are input, and the input field is input. A filter for generating an interpolated pixel of the field to be subjected to the progressive scan conversion by applying a filter process to pixels of at least one of the fields; and from the frame memory, the field to be subjected to the progressive scan conversion, A frame containing the progressive scan conversion target field; A difference calculator for measuring a motion amount of the field to be subjected to progressive scan conversion by inputting a field to be subjected to sequential scan conversion or a frame including the field to be subjected to progressive scan conversion and an adjacent field or frame, and calculating a difference therebetween. And, based on the amount of motion measured by the difference calculator, synthesizes a filter coefficient setting device that changes the filter characteristics of the filter device, the interlaced scan image, and an interpolation pixel generated by the filter device, A double-speed converter for generating a progressively scanned image.

According to the progressive scan conversion method of the present invention, a decoding process is performed on a code string obtained by coding an interlaced scan image composed of a plurality of fields using motion compensation on a field-by-field or frame-by-frame basis. A progressive scan conversion method for converting the interlaced scan image obtained by the decoding process into a progressive scan image, the decoding process being performed on the interlaced scan image to obtain a decoded image; A decoding step of obtaining a motion vector pointing to a default reference field for the target field at the time; A motion vector conversion step for converting into a motion vector of a size corresponding to a unit time interval Acquiring pixels based on the motion vector converted by the motion vector conversion step from a reference field that is a field before and after the field to be subjected to the progressive scan conversion processing, An inter-field interpolation pixel generation step of generating one interpolation pixel, and an intra-field interpolation pixel generation step of generating a second interpolation pixel using a pixel in the progressive scan conversion target field,
A weighting factor determining step of determining a weighting factor indicating a weighting ratio between the first interpolation pixel and the second interpolation pixel; and the first interpolation pixel and the second interpolation pixel Generating a third interpolated pixel by weighting and averaging by a coefficient, and generating a progressively scanned image by interpolating the decoded image using the third interpolated pixel. Features.

In the progressive scanning conversion method according to the twelfth aspect of the present invention, a decoding process is performed on a code sequence obtained by coding an interlaced scanning image composed of a plurality of fields using motion compensation on a field-by-field or frame-by-frame basis. And performing a decoding process on the interlaced scan image to obtain a decoded image by performing a decoding process on the interlaced scan image. A decoding step of obtaining a motion vector pointing to a default reference field for the target field, and a motion vector having a magnitude corresponding to a time interval from the target field to the default reference field corresponding to each field. , A motion vector conversion step for converting to a motion vector of a size corresponding to And a motion vector determining step of determining the validity of the motion vector converted by the motion vector converting step; and a reference field being a field before and after the field to be subjected to the sequential scan conversion processing. An inter-field interpolation pixel generation step of obtaining a pixel based on the motion vector converted by the conversion step and the determination result in the motion vector determination step, and generating a first interpolation pixel for the progressive scan conversion target field; An intra-field interpolation pixel generation step of generating a second interpolation pixel using a pixel in the progressive scan conversion target field,
A weighting factor determining step of determining a weighting factor indicating a weighting ratio between the first interpolation pixel and the second interpolation pixel; and the first interpolation pixel and the second interpolation pixel Generating a third interpolated pixel by weighting and averaging by a coefficient, and generating a progressively scanned image by interpolating the decoded image using the third interpolated pixel. Features.

According to a thirteenth aspect of the present invention, in the progressive scanning conversion method according to the eleventh or twelfth aspect, the time interval of a certain unit in the motion vector conversion step is as follows: It is characterized by a time interval corresponding to one field.

According to a fourteenth aspect of the present invention, there is provided the progressive scan conversion method according to the eleventh or twelfth aspect, wherein the inter-field interpolation pixel generating step, the weight coefficient determining step, In addition, the processing of the progressive scanning image generation step is performed in a unit smaller than an image unit accompanied by a motion vector at the time of the motion compensation.

In the progressive scan conversion method according to a fifteenth aspect of the present invention, in the progressive scan conversion method according to the eleventh or twelfth aspect, the code sequence is a code sequence encoded by an MPEG system. It is characterized by being.

According to a sixteenth aspect of the present invention, in the progressive scanning conversion method according to the eleventh or twelfth aspect, the motion vector conversion step is performed between one line of a frame structure. When the distance is one pixel, the vertical component of the motion vector is converted into an even value.

According to a seventeenth aspect of the present invention, in the progressive scanning conversion method according to the twelfth aspect, the motion vector determining step includes the step of: Is determined to be valid if the magnitude of the is less than or equal to a predetermined value.

In the progressive scanning conversion method according to a twelfth aspect of the present invention, in the progressive scanning conversion method according to the twelfth aspect, the step of determining a motion vector may be such that a distance between one line of a frame structure is one pixel. In the above case, among the motion vectors converted in the motion vector conversion step, a motion vector whose vertical component is an even value is determined to be valid.

According to a nineteenth aspect of the present invention, in the progressive scanning conversion method of the eleventh aspect, the inter-field interpolation pixel generation step is performed by the motion vector conversion step. Calculating an evaluation scale for selecting a motion vector optimal for generating the first interpolation pixel using a motion vector, and generating the first interpolation pixel using a motion vector having the best evaluation scale It is characterized by the following.

According to a twentieth aspect of the present invention, in the progressive scanning conversion method of the eleventh aspect, the inter-field interpolation pixel generation step is performed by the motion vector conversion step. Using a motion vector and a motion vector in the opposite direction to the motion vector, an evaluation scale for selecting a motion vector optimal for generating the first interpolation pixel is calculated, and the motion in which the evaluation scale is the best is calculated. Generating the first interpolated pixel using a vector, wherein the reverse motion vector has a direction opposite to the direction of the motion vector converted by the motion vector conversion step, and refers to the motion vector. The field is a motion vector pointing to a reference field whose context is opposite to that of the target field. .

According to a twenty-first aspect of the present invention, in the progressive scanning conversion method according to the twelfth aspect, the inter-field interpolation pixel generation step is performed by the motion vector conversion step. Among the motion vectors, an evaluation scale for selecting an optimal motion vector for generating the first interpolation pixel is calculated using the motion vector determined to be valid in the motion vector determination step, and the evaluation scale is The first interpolation pixel is generated using a best motion vector.

In the progressive scan conversion method according to a twelfth aspect of the present invention, in the progressive scan conversion method according to the twelfth aspect, the inter-field interpolation pixel generation step is converted by the motion vector conversion step. Of the motion vectors, using an effective motion vector determined to be valid in the motion vector determination step and a motion vector in the opposite direction to the motion vector, a motion vector optimal for generating the first interpolation pixel is calculated. Calculating an evaluation scale for selection, and generating the first interpolated pixel using a motion vector having the best evaluation scale, wherein the reverse motion vector is in a direction different from the effective motion vector. And the reference field whose context is opposite to that of the reference field indicated by the effective motion vector. Characterized in that it is a motion vector pointing to de.

According to a twenty-third aspect of the present invention, in the progressive scanning conversion method according to any one of the nineteenth to twenty-second aspects, the inter-field interpolation pixel generating step includes the step of: Using the motion vector converted by the vector conversion step and the motion vector whose motion is 0, an evaluation scale for selecting a motion vector optimal for generating the first interpolation pixel is calculated, and the evaluation scale is The first interpolation pixel is generated using a best motion vector.

According to a twenty-fourth aspect of the present invention, in the progressive scanning conversion method according to any one of the nineteenth to twenty-second aspects, the evaluation scale is determined by the motion vector conversion step. It is a sum of pixel difference absolute values of a pixel of a reference field indicated by the converted motion vector and the second interpolation pixel.

In the progressive scanning conversion method according to a twenty-fifth aspect of the present invention, in the progressive scanning conversion method according to the twenty-third aspect, the evaluation scale indicates a motion vector converted by the motion vector converting step. It is a sum of pixel difference absolute values of a pixel of a reference field and the second interpolation pixel.

Further, in the progressive scan conversion method according to claim 26 of the present invention, in the progressive scan conversion method according to claim 20 or 22, the evaluation scale is converted by the motion vector conversion step. It is a sum of pixel difference absolute values of a pixel in the reference field indicated by the motion vector and a pixel in the reference field indicated by the reverse motion vector.

According to a twenty-seventh aspect of the present invention, in the progressive scanning conversion method according to the twentieth or twenty-second aspect, the inter-field interpolation pixel generation step includes the motion vector conversion step. Is used to calculate an evaluation scale for selecting a motion vector that is optimal for generating the first interpolation pixel, using the motion vector whose motion is 0, and the evaluation scale is the best. Generating the first interpolation pixel using a motion vector, wherein the evaluation scale is a pixel in a reference field indicated by the motion vector converted by the motion vector conversion step, and the motion vector in the reverse direction indicates It is a sum of pixel difference absolute values with the pixel of the reference field.

In the progressive scanning conversion method according to a twenty-eighth aspect of the present invention, an interpolated pixel is generated for an interlaced scan image composed of a plurality of fields by using pixels in each of the fields to convert the interlaced scan image. In a progressive scan conversion method for converting to a progressive scan image, an edge detection step of detecting, as an edge direction, a direction indicated by a straight line that passes through an interpolation position for generating the interpolation pixel and connects peripheral pixels of the interpolation position, Edge reliability determining step of obtaining the strength of correlation between pixels existing in the direction of the edge as the reliability of the edge, and using the pixel existing in the direction of the edge when the reliability of the edge is a predetermined value or more. When the reliability of the edge is less than a predetermined value, an interpolated pixel is generated by using a pixel existing in the vertical direction of the interpolation position. Characterized in that it comprises a between pixel generating step.

According to a twenty-sixth aspect of the present invention, in the progressive scanning conversion method according to the eleventh or twelfth aspect, the step of generating an intra-field interpolation pixel includes the step of: An edge detection step of detecting, as an edge direction, a direction indicated by a straight line connecting pixels surrounding the interpolation position, passing through an interpolation position for generating a pixel, and determining an edge strength of a correlation between pixels existing in the edge direction. Edge reliability determination step to determine the reliability of the edge, when the reliability of the edge is a predetermined value or more, the second interpolation pixel is generated using a pixel existing in the direction of the edge, the edge of the edge, If the reliability is less than a predetermined value, an interpolation pixel generation step of generating the second interpolation pixel using pixels existing in the vertical direction of the interpolation position,
It is characterized by including.

In the progressive scanning conversion method according to a thirty-second aspect of the present invention, in the progressive scanning conversion method according to the twenty-eighth aspect, the step of determining the edge reliability includes the step of determining whether or not a pixel exists in a direction of the edge. If the inter-difference value is smaller than the inter-pixel difference value of the pixels existing in the vertical direction of the interpolation position, it is determined that the reliability of the edge is equal to or more than a predetermined value.

In the progressive scan conversion method according to a thirty-first aspect of the present invention, in the progressive scan conversion method according to the thirty-ninth aspect, the step of determining the edge reliability includes the step of determining a pixel existing in a direction of the edge. If the inter-difference value is smaller than the inter-pixel difference value of a pixel existing in the vertical direction of the interpolation position, it is determined that the reliability of the edge is equal to or more than a predetermined value.

In the progressive scanning conversion method according to a thirty-second aspect of the present invention, in the progressive scanning conversion method according to the twenty-eighth aspect, the edge reliability determining step uses a pixel existing in a direction of the edge. If the calculated interpolated pixel value is a value between the pixel values of the pixels above and below the interpolation position, it is determined that the reliability of the edge is equal to or greater than a predetermined value.

In the progressive scan conversion method according to a thirty-third aspect of the present invention, in the progressive scan conversion method according to the thirty-ninth aspect, the edge reliability determining step uses a pixel present in a direction of the edge. If the obtained interpolated pixel value is a value between the pixel values of the pixels above and below the interpolation position, it is determined that the reliability of the edge is equal to or more than a predetermined value.

Further, according to a progressive scan conversion method described in claim 34 of the present invention, in the progressive scan conversion method according to claim 11 or 12, the intra-coding on the field to be progressively converted is performed. For the progressive scan conversion target image region, a peripheral image region located around the progressive scan conversion target image region, or the progressive scan conversion target image region in a frame immediately before or immediately after the progressive scan conversion target field. The sequential scan conversion processing is performed using a motion vector attached to an image area at the same position as the above.

In the progressive scanning conversion method according to claim 35 of the present invention, in the progressive scanning conversion method according to claim 11 or 12, the code string decoded in the decoding step is recorded. In the case of a code string recorded on a medium and read in a fast-forward or fast-reverse mode, the decoded image is interpolated using only the second interpolation pixels generated by the intra-field interpolation pixel generation step. , Generating a progressively scanned image.

Further, the progressive scan conversion apparatus according to claim 36 of the present invention performs decoding processing on a code sequence obtained by coding an interlaced scan image composed of a plurality of fields using motion compensation on a field-by-field or frame-by-frame basis. And performing a decoding process on the interlaced scan image to obtain a decoded image by performing a decoding process on the interlaced scan image. At the time, a decoder for obtaining a motion vector pointing to a predetermined reference field for the target field, an image memory for storing the decoded image, a parameter memory for storing the motion vector, and each of the parameters read from the parameter memory. The time interval in the field from the target field to the default reference field A motion vector converter for converting a motion vector of a corresponding size into a motion vector of a size corresponding to a fixed unit of time interval, and a field before and after a field for progressive scan conversion to be subjected to progressive scan conversion processing. An inter-field interpolation pixel generator for obtaining a pixel from a reference field based on the motion vector converted by the motion vector converter and generating a first interpolation pixel for the field to be subjected to the progressive scan conversion; An intra-field interpolation pixel generator that generates a second interpolation pixel using a pixel in a target field, and a weight coefficient indicating a weight ratio between the first interpolation pixel and the second interpolation pixel are determined. A weight coefficient determiner, the first interpolation pixel and the second interpolation pixel are weighted and averaged by the weight coefficient. A progressively scanned image generator that generates progressively scanned images by generating three interpolated pixels and interpolating the decoded image read from the image memory using the third interpolated pixels, I do.

A progressive scanning conversion apparatus according to claim 37 of the present invention performs decoding processing on a code sequence obtained by coding an interlaced scan image composed of a plurality of fields using motion compensation on a field-by-field or frame-by-frame basis. And performing a decoding process on the interlaced scan image to obtain a decoded image by performing a decoding process on the interlaced scan image. Sometimes used, a decoder for obtaining a motion vector pointing to a predetermined reference field for the target field, an image memory for storing the decoded image, a parameter memory for storing the motion vector, and read from the parameter memory. In each of the fields, from the target field to the predetermined reference field A motion vector converter for converting a motion vector having a size corresponding to a time interval into a motion vector having a size corresponding to a certain unit of time interval, and determining the validity of the motion vector converted by the motion vector converter; And a reference vector that is a field before and after the field to be subjected to the progressive scan conversion to be subjected to the progressive scan conversion processing, from the motion vector converted by the motion vector converter to the determination result by the motion vector determiner. An inter-field interpolation pixel generator that obtains a pixel based on the field and generates a first interpolation pixel for the field of the progressive scan conversion, and reads a pixel in the field of the progressive scan conversion to generate a second interpolation pixel In-field interpolation pixel generator, between the first interpolation pixel and the second interpolation pixel A weight coefficient determiner for determining a weighting factor indicating a weighting ratio to generate said first interpolated pixel, a third interpolated pixel by the said second interpolation pixel weighted average by the weighting factor,
A progressive scanning image generator that generates a progressive scanning image by interpolating the decoded image read from the image memory using the third interpolation pixel.

[0062]

Embodiments of the present invention will be described below with reference to the drawings. (Embodiment 1) Hereinafter, Embodiment 1 of the present invention will be described with reference to FIG.
This will be described with reference to FIG. FIG. 1 shows a frame memory 101.
, A VT filter device 102, and a filter coefficient setting device 10
3 is a block diagram of a progressive-scan conversion device including a differential operation unit, a differential operation unit, and a double-speed converter. Also, FIG.
FIG. 4 is a schematic diagram illustrating a state of a time-vertical plane of an interlaced image, in which white circles indicate pixels (scanning lines), and vertical arrangement of white circles indicates pixels belonging to the same field.

First, the progressive scan converter stores the input interlaced image in the frame memory 101. Here, a case will be described in which four fields # n−2 to # n + 1 of FIG. 2 are accumulated, and field #n is sequentially scan-converted. In this case, the VT filter 102
Data of fields # n-1, #n, and # n + 1 are input. Then, in FIG. 2, when an interpolation pixel at the position k is generated, a filter process is performed on an adjacent pixel at the position k.
Here, the case where the filtering process is performed on the pixels a to j will be described.

The difference calculator 104 outputs data for two frames including the field #n to be progressively converted, and the frames # m-1, #m (fields # n-2 to #m).
n + 1). The difference calculator 104 calculates the difference between the frames # m-1 and #m in order to obtain the motion amount of the frame #n. For this difference, a difference value between pixels corresponding to the positions of both frames is calculated, and the absolute value sum of the difference values is calculated as a motion amount. Then, the result is output to the filter coefficient setting unit 103.

The filter coefficient setting unit 103 determines a filter coefficient based on the value of the sum of absolute differences input from the difference calculator 104. For example, threshold values TH1, TH
2. Set TH3 (though TH1 <TH2 <TH3), and if the sum of absolute differences is smaller than the threshold value TH1, a still image;
If it is H2 or more and less than TH3, it is determined to be a quasi-moving image, and if it is TH3 or more, it is determined to be a moving image. Then, a predetermined filter coefficient corresponding to each of the still image, the quasi-still image, the quasi-moving image, and the moving image is output to the VT filter 102.

Here, basically, the VT filter 102 extracts a low frequency component in the vertical direction from the field to be sequentially converted by the filter coefficient setting unit 103.
A coefficient is set so as to extract a high frequency component in the vertical direction from the adjacent field. For example, when generating the interpolated pixel k shown in FIG. 2, the VT filter 102 outputs the vertical high frequency components from the pixels a to c and the pixels h to j, and the vertical high frequency components from the pixels d to g. Are set so as to extract the low-frequency components of.

Further, the filter coefficient setting unit 10
No. 3 sets a filter coefficient so that a component (gain) from an adjacent field becomes smaller as the image becomes a quasi-still image, a quasi-moving image, and a moving image. For example, in the case of a moving image, a component (gain) from an adjacent field may be set to 0.
FIG. 3 shows an example of the frequency characteristics of the VT filter 102 in this case. 3A and 3B show gains for the vertical frequency in the filter. FIG. 3A is for a still image, FIG. 3B is for a quasi-still image, FIG.
FIG. 3D shows a moving image. Here, in FIG. 3, a solid line indicates a filter characteristic for a field to be subjected to progressive scan conversion, and a broken line indicates a frequency characteristic for an adjacent field.

The VT filter 102 applies a filter to an input pixel using the filter coefficient value input from the filter coefficient setting unit 103 to generate an interpolated pixel and outputs the interpolated pixel. The generated interpolation pixel is input to the double speed converter 105. The double speed converter 105 includes a frame memory 101
, And pixel data of the progressive scan conversion field, and interpolated pixel data from the VT filter 102. The double-speed converter 105 combines these data, changes the frame rate to double, and outputs the result as a progressive image.

As described above, in the first embodiment, when converting an interlaced image into a progressive image,
The T filter 102 applies a VT filter to the field to be subjected to the progressive scan conversion and the pixels in the fields before and after the field to generate an interpolation pixel. At that time,
The difference calculator 104 obtains the sum of absolute differences between pixels between the frame including the field to be subjected to the progressive scan conversion and the immediately preceding frame, and the filter coefficient setting unit 103 calculates The coefficients of the VT filter are determined. The coefficient was determined to be a moving image as the sum of absolute differences was larger, and was determined so that the contribution (gain) from the adjacent field was reduced.

Therefore, in the case of a still image, a high-resolution progressive image can be obtained as in the case of the conventional VT filter system, and in the case of a moving image, a portion of the image where the image quality has deteriorated in the conventional VT filter can be obtained. Image quality can be greatly improved. Further, since these operations can be realized by one filter, cost can be reduced.

In the first embodiment, the interlaced image is a still image, a quasi-still image, a quasi-moving image, and a moving image based on the absolute value of the difference input from the difference calculator 104 by the filter coefficient setting unit 103. Is determined, and the filter coefficient is determined in four stages. Therefore, compared with the conventional progressive scan conversion method in which the filter coefficient is switched in two stages depending on whether the interlaced image is a still image or a moving image, it is possible to generate an interpolated pixel that more matches the motion of the image. .

In the first embodiment, when the motion amount of the field to be sequentially converted is calculated by the difference calculator 104, the sum of absolute differences between the frame including the field to be sequentially converted and the previous frame is calculated. May be calculated, but the sum of the absolute values of the differences between the field to be sequentially converted and the preceding field may be calculated. Further, the sum of absolute differences between the frame including the field to be subjected to the progressive scan conversion and the subsequent frame or the field to be subjected to the progressive scan conversion and the subsequent field may be obtained. Further, the larger value or the average value of the sum of absolute differences between the frame including the field for progressive scan conversion and the preceding and succeeding frames or between the field for progressive scan conversion and the preceding and succeeding fields may be obtained.

In the first embodiment, a case has been described in which the VT filter 102 generates an interpolation pixel by applying a filter using ten pixels. It may be.

In the first embodiment, the VT filter 102 controls the fields # n-1, #n, # n +
The interpolated pixel of the field #n is generated by using the pixels of the three fields of 1, which are the fields # n-1 and #n,
Alternatively, an interpolated pixel may be generated using pixels of two fields such as fields #n and # n + 1. Note that since the frame memory 101 stores the interlaced scanning image in field units, in this case, the size of the frame memory can be reduced as compared with the case where the interpolation pixels are generated using the pixels of three fields. It is possible to reduce the amount of processing at the time of filter processing.

In the first embodiment, the case where the filter coefficient setting unit 103 determines the amount of motion in four steps has been described, but the number of steps may be another value. Also, in the first embodiment, the case has been described where the motion amount is determined in units of fields or frames, but this is done by dividing the screen into several regions,
The amount of motion may be detected for each of the divided regions, and the filter coefficient may be determined.

Embodiment 2 Hereinafter, Embodiment 2 of the present invention will be described with reference to FIG. FIG. 4 shows a frame memory 101, a VT filter 102, a filter coefficient setter 103, a difference calculator 404, and a double speed converter 1
FIG. 5 is a block diagram of a progressive scan conversion device provided with the device No. 05.

First, the progressive scan converter stores the input interlaced image in the frame memory 101. Here, a case will be described in which three fields # n-1 to # n + 1 shown in FIG. 2 are accumulated, and field #n is sequentially scan-converted.

In this case, the VT filter 102 inputs the data of the fields # n-1, #n, # n + 1.
Then, in FIG. 2, when generating an interpolation pixel at the position k, for example, a filter process is performed on pixels a to j that are adjacent pixels.

The difference calculator 404 receives the pixels b and i of the adjacent field at the same position as the interpolation position among the pixels used for the filtering by the VT filter 102.
The difference calculator 404 calculates the difference absolute value of these pixel values. Then, the result is used as the filter coefficient setting unit 103
Output to

The filter coefficient setting unit 103 determines a filter coefficient based on the absolute value of the difference input from the difference calculator 404. For example, threshold values TH1, TH2,
TH3 (though TH1 <TH2 <TH3) is set, and if the absolute value of the difference is smaller than the threshold value TH1, a still image is displayed. , TH3 or more, it is determined to be a moving image. Then, a predetermined filter coefficient corresponding to each of the still image, the quasi-still image, the quasi-moving image, and the moving image is output to the VT filter 102. As described above, in the second embodiment, since the difference absolute value between pixels in adjacent fields at the same position as the interpolation position is obtained by the difference calculator 404, the filter coefficient setting unit 103 calculates the amount of motion for each pixel. This makes it possible to determine, and when there is an object or the like moving in the screen, it is possible to prevent the image quality degradation caused by the conventional VT filter for the object. The characteristics of the coefficients of the VT filter 102 are the same as in the first embodiment, and a description thereof will not be repeated.

The VT filter 102 applies a filter to an input pixel using the filter coefficient value input from the filter coefficient setting unit 103 to generate and output an interpolation pixel. The generated interpolation pixel is input to the double speed converter 105. The double speed converter 105 includes a frame memory 101
, And pixel data of the progressive scan conversion field, and interpolated pixel data from the VT filter 102. The double-speed converter 105 combines these data, changes the frame rate to double, and outputs the result as a progressive image.

As described above, according to the second embodiment, when the interlaced image is converted into the progressive image, the VT
The filter 102 generates an interpolated pixel by applying a VT filter to pixels in a field to be subjected to progressive scan conversion and pixels in fields before and after the field. At that time,
The difference calculator 404 calculates the absolute value of the difference between the pixels in the adjacent fields located at the same position as the interpolation position, and based on the value of the absolute value, the filter coefficient setting unit 103 determines the coefficient of the VT filter. I made it. The coefficient is determined to be a moving image as the difference absolute value is larger, and is determined so as to reduce the contribution from the adjacent field.

Therefore, in the case of a still image, a high-resolution progressive image can be obtained by the conventional VT filter, and when there is an object moving on the screen,
For the object, it is possible to prevent the image quality from deteriorating due to the conventional VT filter. In addition, these operations are 1
Since it can be realized with one filter, cost can be reduced.

In the second embodiment, the case where the difference value between the pixels b and i is calculated by the difference calculator 404 has been described, but this may be the difference value of another pixel. For example, the sum of absolute differences between pixels corresponding to the pixels a to c and the pixels h to j may be calculated, or the absolute difference between the average value of the pixels e and f and the pixel b may be calculated. Further, in the second embodiment, a case has been described where the VT filter 102 generates an interpolation pixel by applying a filter using ten pixels. It may be a value.

Also, in the second embodiment, the VT filter 102 determines whether the fields # n-1, #n, # n +
The interpolated pixel of the field #n is generated by using the pixels of the three fields of 1, which are the fields # n-1 and #n,
Alternatively, an interpolated pixel may be generated using pixels of two fields such as fields #n and # n + 1. Note that since the frame memory 101 stores the interlaced scanning image in field units, in this case, the size of the frame memory can be reduced as compared with the case where the interpolation pixels are generated using the pixels of three fields. It is possible to reduce the amount of processing at the time of filter processing. Further, in the second embodiment, the case has been described where the amount of motion is determined in four steps in the filter coefficient setting unit 103, but the number of steps may be another value.

Embodiment 3 Hereinafter, Embodiment 3 of the present invention will be described. FIG. 5 is a block diagram showing a configuration of a progressive scan conversion device including an MPEG video decoder 501 and a progressive scan converter 508. The progressive scan converter receives MPEG video obtained by motion-compensating and predictively encoding an interlaced image by the MPEG method. The MPEG video decoder 501
Variable length decoder 502, inverse quantizer 503, inverse DC
A T unit 504, an adder 505, a motion compensator 506,
An image memory 507 is provided. FIG. 6 is a block diagram showing the configuration of the progressive scan converter 508, which includes a parameter memory 601, a motion vector converter 602, a motion vector determiner 603, an inter-field interpolation pixel generator 604, an intra-field An interpolation pixel generator 605, a weight coefficient determiner 606, and a progressively scanned image generator 607 are provided.

The MPEG video code string input to the MPEG video decoder 501 is subjected to variable length decoding by the variable length decoder 502. DCT coefficients and the like obtained by performing variable length decoding on the MPEG video code string are as follows:
Output to the inverse quantizer 503. The variable-length decoder 502 obtains information such as a motion vector, a picture type, a macro block type, a temporal reference, and a motion vertical field select (MVFS) by performing variable-length decoding on the MPEG video. To the motion compensator 506 and the progressive scan converter 5
08 and is output. Here, the picture type is information indicating whether the decoded image data is an I picture, a P picture, or a B picture. The macroblock type refers to a macroblock having a size of, for example, 16 × 16 pixels in decoded image data in any prediction mode (no prediction, forward prediction, backward prediction, bidirectional prediction, and the like). This is information indicating whether encoding has been performed. The temporal reference is information indicating a frame number of a picture. MVFS is information indicating whether a reference image in field prediction is a top field image or a bottom field image when an interlaced image is encoded by a frame structure and field prediction is used as motion compensation prediction. .

The inverse quantizer 503 performs inverse quantization on the input DCT coefficient. And the inverse DCT unit 504
Performs inverse DCT on the result, generates image data, and outputs the image data to the adder 505.

The motion compensator 506 reads the reference image data from the image memory 507 and performs motion compensation. Note that the decoded image data output from the adder 505 is stored in the image memory 507 as described later. Then, the decoded image data becomes reference image data when performing motion compensation. For example, in the MPEG2 encoding method,
The image data is motion compensated in units of 16 × 16 pixels (macroblock). Therefore, the input code string is MPEG2
In the case of a code string encoded by the encoding method of
The reference image data to be motion-compensated by the motion compensator 506 refers to decoded 16 × 16 pixel data (macroblock). Note that the motion compensator 506 does not operate when the decoded macroblock is an intra (intra-frame coded) macroblock.
If the decoded macroblock is a non-intra macroblock, motion compensation is performed based on the motion vector input from the variable length decoder 502, and the motion-compensated image data is output to the adder 505. .

In the adder 505, when the decoded macro block is an intra macro block, the inverse DCT
The image data input from the device 504 is output without any processing. On the other hand, if the decoded macroblock is a non-intra macroblock, the image data input from the inverse DCT unit 504 and the motion compensator 506
, And the result is output. The image data output from the adder 505 is stored in the image memory 507 as decoded image data.

The progressive scan converter 508 is connected to the image memory 50.
7, the decoded image data is read out in time order and processed. This operation will be described below. Progressive scan converter 508
, Information such as a motion vector, a picture type, a temporal reference, a macro block type, and MVFS are input from the variable length decoder 502 to the image memory 507.
Receives decoded image data. Variable length decoder 5
02, motion vector, picture type, macro block type, temporal reference, MVF
Information such as S is input to the parameter memory 601 and temporarily stored. This is the MPEG video decoder 501
Performs the processing of each frame in the order of the code string, whereas the progressive scan converter 508 performs the processing in the order of time, so that the time difference is compensated.

The motion vector converter 602 reads information such as a motion vector, a picture type, a temporal reference, and an MVFS from the parameter memory 601.
The magnitude of the motion vector is converted. Here, the motion vector converter 602 uses the information of the picture type, the temporal reference, and the MVFS to determine the field to be subjected to the progressive scan conversion and to refer to the field to be subjected to the progressive scan conversion at the time of motion compensation for encoding. It is possible to obtain information indicating how many fields are separated from the reference field.

The operation of converting a motion vector will be described below with reference to FIG. FIG. 7 is a schematic diagram showing a state of a decoded image (interlaced image). The white circles indicate the arrangement of scanning lines (pixels) viewed from the horizontal direction of the screen, where the vertical direction is the vertical direction and the horizontal direction is the time direction. That is, white circles vertically arranged in a line indicate scanning lines (pixels) belonging to the same field.
Is shown. Now, a case in which the motion vector of the macroblock a in the field for progressive scan conversion is converted will be described. Assuming that the motion vector of the macro block a is the motion vector A, the macro block a refers to the reference area b in the reference field three frames (six fields) apart at the time of encoding. The motion vector converter 602 converts the motion vector A into a one-field unit motion vector. That is, the motion vector A is converted into a motion vector B indicated by a broken line. Then, the converted motion vector is output to the motion vector determiner 603. In FIG. 7,
Although only the movement in the vertical direction is shown, the same processing is performed for the movement in the horizontal direction. FIG. 7 shows only the motion vector in the forward direction,
In the case where there is a backward motion vector, the motion vector converter 602 performs a similar conversion process on the backward motion vector.

The motion vector determiner 603 determines the validity of the motion vector output from the motion vector converter 602. FIG. 8 shows a flowchart of the determination procedure. First, in step S801, it is determined whether the magnitude of the motion vector is equal to or smaller than a threshold. The determination here may be “Yes” if both the horizontal and vertical motion vectors are equal to or smaller than the threshold, or may be “Yes”.
If the sum of the squares of the vertical motion vectors is equal to or less than the threshold value, “Yes” may be set. If the determination result in step S801 is “Yes”, in step S802,
It is determined whether the vertical motion vector value is an even number. Here, the unit of the vertical movement is one pixel of the distance between one line in the frame structure. For example, the vertical size of the motion vector B in FIG. 7 is 2 pixels. The reason why the motion vector in the vertical direction is even in step S802 is selected because an interpolation value cannot be directly obtained from the reference field if the motion vector is odd. If the determination result of step S801 is “No”,
In step S804, the motion vector is determined to be invalid, and is not used in the subsequent processing. If the determination result in step S802 is “Yes”, in step S803,
A motion vector that satisfies the conditions of steps S801 and S802 is determined as a valid motion vector, and is used in the subsequent processing. Also, the determination result of step S802 is “N
If it is “o”, the motion vector is determined to be invalid in step S804, and is not used in the subsequent processing. Here, when there are a plurality of motion vectors (forward and backward) in the macroblock, The determination result and the motion vector by the motion vector determiner 603 are output to the inter-field interpolation pixel generator 604.

The inter-field interpolation pixel generator 604 obtains pixels from the reference field based on the motion vector input from the motion vector determiner 603 and the result of the determination, and generates inter-field interpolation pixels for the field to be sequentially subjected to scan conversion. . Subsequent processing is performed in macroblocks or in units smaller than macroblocks. In the following description, it is assumed that processing is performed in units of 8 horizontal pixels and 4 vertical pixels (hereinafter, referred to as blocks).
In the following, when simply referred to as “motion vector”, it refers to a motion vector that has been converted by the motion vector converter 602.

FIGS. 9 and 10 show an operation example of the inter-field interpolation pixel generator 604. FIG. First, when there are a plurality of valid motion vectors, the inter-field interpolation pixel generator 604 calculates an evaluation scale for selecting which motion vector is used to generate an inter-field interpolation pixel.
In the present embodiment, a calculation method for obtaining two types of evaluation scales will be described. FIG. 9 is a flowchart showing an operation example of the inter-field interpolation pixel generator 604 when selecting a motion vector according to the first evaluation scale, and FIG. 10 is a flowchart when selecting a motion vector according to the second evaluation scale. 6 is a flowchart showing an operation example of the inter-field interpolation pixel generator 604 of FIG.

Hereinafter, a case where a motion vector is selected according to the first evaluation scale will be described with reference to FIG. First,
In step S901, it is determined whether or not a valid forward motion vector exists in a block to be subjected to progressive scan conversion. Here, the case where there is a valid forward motion vector means that the macroblock including the block to be sequentially converted has the forward motion vector, and the motion vector determiner 603 determines that the motion vector is valid. This is when it is determined. If the result of the determination in step S901 is “Yes”, the flow proceeds to step S902. Also, “N
If it is o ", the process proceeds to step S903. In step S902, an evaluation scale is calculated using the forward motion vector.

Hereinafter, a method of calculating the first evaluation scale using the forward motion vector will be described with reference to FIG.
FIG. 13 is a schematic diagram showing a state of a decoded image (interlaced image), similarly to FIG. Now, it is assumed that an inter-field interpolation pixel is generated for the block c of the field to be sequentially converted and that the forward motion vector of the macroblock to which the block c belongs is the motion vector C. In this case, the sum of absolute differences between the block d in the forward reference field indicated by the motion vector C and the block e in the backward reference field indicated by the motion vector C ′ in the reverse direction of the motion vector C is calculated. Note that the motion vector C ′ in the reverse direction is a motion vector indicating a reference field whose direction is opposite to that of the motion vector C, and which is in reverse relation to the reference field indicated by the motion vector C. Means that. Also,
The calculation of the sum of absolute differences between blocks is performed by taking the difference between the pixels connected by arrows of blocks d and e,
This means that the sum of those absolute differences is taken in the block. Then, this sum of absolute differences is used as a first evaluation scale. Although FIG. 13 shows only the pixels in the vertical direction, the pixels in the horizontal direction are also targets for calculating the sum of absolute differences between pixels as a first evaluation scale.

In step S903, it is determined whether a valid backward motion vector exists in block c.
If the result of the determination in step S903 is "Yes", the flow proceeds to step S904. If “No”, the process proceeds to step S905.

In step S904, an evaluation scale is calculated using the backward motion vector. The method for calculating the evaluation scale is the same as that in step S902, and a description thereof will not be repeated. However, the difference is that the motion vector used is a backward motion vector.

Finally, in step S905, based on the evaluation scale obtained by the above processing, an interpolation method, that is, which motion vector is used to generate inter-field interpolation pixels is selected, and inter-field interpolation pixels are generated. I do. The selection method will be described later.

Next, a case where a motion vector is selected based on the second evaluation scale will be described with reference to FIG. First, in step S1001, it is determined whether a valid forward motion vector exists in a block to be subjected to progressive scan conversion. The determination result of step S1001 is “Yes”
If it is, the process proceeds to step S1002. If “No”, the process proceeds to step S1006. In step S1002, an evaluation scale is calculated using the forward motion vector. Hereinafter, a second evaluation scale calculation method will be described with reference to FIG. FIG. 14 is a schematic diagram showing a state of a decoded image (interlaced image), similarly to FIG. Now, it is assumed that an inter-field interpolation pixel is generated for the block c, and that the macroblock to which the block c of the field to be sequentially converted belongs has only the forward motion vector C as a valid motion vector. In this case, the pixels of the block d in the forward reference field indicated by the motion vector C and the block c generated by the intra-field interpolation pixel generator 605 (the generation method will be described later)
Calculate the sum of the absolute values of the differences with the interpolation pixel in Here, the calculation of the sum of absolute differences between the blocks means that the differences between the pixels connected by the arrows of the blocks d and c are obtained, and the sums of the absolute values in the block are obtained. Then, this sum of absolute differences is used as a second evaluation scale. Figure 1
In FIG. 4, only the pixels in the vertical direction are shown, but the pixels in the horizontal direction are also targets for calculating the sum of absolute differences between pixels as a second evaluation scale.

Another method of calculating the second evaluation scale will be described with reference to FIG. Now, it is assumed that an inter-field interpolation pixel is generated for the block c, and the macroblock to which the block c belongs has only the forward motion vector C as a valid motion vector. In this case, the pixel of the block e in the backward reference field indicated by the motion vector C ′ in the direction opposite to the motion vector C and the intra-field interpolation pixel in the block c generated by the intra-field interpolation pixel generator 605 described later. Calculate the sum of absolute differences.
Here, the calculation of the sum of the absolute differences between the blocks means that the differences between the pixels connected by the arrows of the blocks c and e are obtained, and the sum of the absolute values in the block is obtained. Then, this sum of absolute differences is used as a second evaluation scale. Although FIG. 15 shows only pixels in the vertical direction, the pixels in the horizontal direction are also targets for calculating the sum of absolute differences between pixels as a second evaluation scale.

As another calculation method of the second evaluation scale, as shown in FIG. 15, a pixel block generated by an average value of the pixels corresponding to the positions of the blocks d and e and an intra-field interpolation pixel to be described later are used. A method is also conceivable in which the sum of absolute differences with the intra-field interpolation pixels in the block c generated by the generator 605 is calculated, and the sum of the absolute differences is used as a second evaluation scale.

Subsequently, in step S1003, after calculating the evaluation scale of the effective forward motion vector, it is further determined whether or not an effective backward motion vector exists in the block to be sequentially subjected to scan conversion. If the determination result of step S1003 is “Yes”, step S100
Proceed to 4. If “No”, step S
Proceed to 1008.

In step S1004, an evaluation scale is calculated using the backward motion vector. The method for calculating the evaluation scale is the same as that in step S1002, and a description thereof will not be repeated. However, the difference is that the motion vector used is a backward motion vector.

In step S1005, an evaluation scale is calculated using the bidirectional motion vector. A method of calculating the evaluation scale will be described with reference to FIG. Now, it is assumed that an inter-field interpolation pixel is generated for the block c, and the macroblock to which the block c belongs has a forward motion vector C and a backward motion vector D as motion vectors. In this case, first, the average value of the pixel of the block d in the forward reference field indicated by the motion vector C and the pixel of the block e in the backward reference field indicated by the motion vector D is calculated. The sum of absolute differences with the interpolated pixels in the block c generated by the intra-field interpolated pixel generator 605 described later is calculated. Here, the calculation of the sum of absolute differences between blocks means the average value of pixels connected by arrows of blocks d and e,
This means that the difference between the interpolated pixel in the field of the block c connected by the arrow is obtained, and the sum of the absolute values of the differences is obtained in the block. Then, this sum of absolute differences is used as a second evaluation scale. Although FIG. 16 shows only pixels in the vertical direction, pixels in the horizontal direction are also subjected to the calculation of the sum of absolute differences between pixels as a second evaluation scale.

In step S1006, when there is no valid forward motion vector, it is determined whether or not there is a valid backward motion vector in the block to be sequentially converted. The determination result of step S1006 is “Yes”
If it is, the process proceeds to step S1007. If “No”, the process proceeds to step S1008. In step S1007, an evaluation scale is calculated using the backward motion vector. The method of calculating the evaluation scale is the same as that in step S1004, and thus the description is omitted.

Finally, in step S1008, based on the evaluation scale obtained by the above processing, an interpolation method, that is, which motion vector is used to obtain an inter-field interpolation pixel from which reference field, is selected. Generate an interpolation pixel. The selection method will be described later.

The operation of generating inter-field interpolation pixels based on the evaluation scale obtained as described above will be described below. First, the processing method in step S905, in which the first evaluation scale is used, will be described. In this processing method, when there are forward and backward motion vectors as valid motion vectors, the motion vector having the smallest first evaluation scale is selected. Then, an inter-field interpolation pixel is generated using the motion vector having the smallest evaluation scale. As an example, a method of generating inter-field interpolation pixels when the evaluation scale of the forward motion vector is small will be described with reference to FIG. In this case, the average value of the block d referenced by the forward motion vector C from the forward reference field and the block e referenced by the motion vector in the opposite direction of the forward motion vector C from the backward reference field is defined as an inter-field interpolation pixel. Generate. Here, the calculation of the average value between the blocks means to take the average value between the pixels connected by the arrows of the block d and the block e in FIG. FIG. 17 shows only the pixels in the vertical direction, but the processing is the same for the pixels in the horizontal direction. In this case, the pixel of the block d or the pixel of the block e may be used as it is as the inter-field interpolation pixel.

Next, a description will be given of a processing method in step S1008 in which the second evaluation scale is used. In this processing method, when there are forward, backward, and bidirectional motion vectors as valid motion vectors, one direction of each motion vector, the opposite direction, of the second evaluation scale obtained for the two directions, A motion vector and an interpolation direction with the smallest evaluation scale are selected. Then, an inter-field interpolation pixel is generated based on the selected motion vector and the interpolation direction. The selection of the interpolation direction is
This means selecting from which reference field in the direction indicated by the motion vector the interpolated pixel is obtained, and here,
One of the directions (one direction, opposite direction, bidirectional) for which the evaluation scale is obtained using each motion vector is the interpolation direction. As an example, when the evaluation scale in the case of obtaining in both directions using the forward motion vector is minimized,
A method of generating inter-field interpolation pixels will be described with reference to FIG. FIG. 17 shows a case where the evaluation scale obtained in the bidirectional direction using the forward motion vector C is minimum. In this case, a block d referred to by the forward motion vector C and a block e referenced by a motion vector C ′ in the opposite direction to the forward motion vector C from the backward reference field.
Is generated as an inter-field interpolation pixel.
Here, the calculation of the average value between the blocks means to take the average value of the pixels connected by the arrows of the block d and the block e. When the evaluation scale obtained in one direction using the forward motion vector C is minimized, an inter-field interpolation pixel is generated from the pixel of the block d and obtained in the opposite direction using the forward motion vector C. When the evaluation scale becomes minimum, an inter-field interpolation pixel is generated from the pixel of the block e. FIG. 17 shows only the pixels in the vertical direction, but the processing is the same for the pixels in the horizontal direction.

It is also possible to generate inter-field interpolation pixels using both the first and second evaluation scales. In this case, for example, a motion vector is selected by a first evaluation scale, and an interpolation direction is determined by a second evaluation scale.
Such a method can be used.

Next, using the motion vector converted by the motion vector converter 602 and a motion vector having zero motion (still motion vector), an optimal motion vector for generating an inter-field interpolation pixel is selected. FIG. 11 and FIG. 12 show an operation example of the inter-field interpolation pixel generator 604 in the case where the evaluation scale is calculated and the inter-field interpolation pixel is generated using the motion vector having the best evaluation scale. FIG. 11 is a flowchart illustrating an operation example of the inter-field interpolation pixel generator 604 when selecting a motion vector according to the first evaluation scale. FIG. 12 is a flowchart illustrating the operation when selecting a motion vector according to the second evaluation scale. 6 is a flowchart showing an operation example of the inter-field interpolation pixel generator 604 of FIG.

First, an example of the operation of the inter-field interpolation pixel generator 604 will be described with reference to FIG. In FIG.
The operations in steps S901 to S904 are the same as those described with reference to FIG. In step S1101, it is determined whether a valid motion vector (forward and backward) includes a still motion vector. If the result of the determination in step S1101 is “Yes”, the flow proceeds to step S1102. If “No”, step S11
Proceed to 03.

In step S1102, the first evaluation scale is calculated using the still motion vector. The calculation method will be described with reference to FIG. FIG. 18 is a schematic diagram showing a state of a decoded image (interlaced image), similarly to FIG. Now, assume that an inter-field interpolation pixel is generated for block c. In this case, the sum of absolute differences between the block f in the forward reference field indicated by the still motion vector E and the block g in the backward reference field indicated by the motion vector E ′ in the opposite direction of the still motion vector E is calculated.
The motion vector E ′ in the opposite direction is a motion vector that indicates a reference field whose direction is opposite to that of the motion vector E and whose context is opposite to that of the target field indicated by the motion vector E. means. Here, the calculation of the sum of absolute differences between blocks means that the block f
And the difference between the pixels connected by the arrow of the block g is obtained, and the sum of the absolute values of the differences is obtained in the block. Then, the sum of absolute differences is used as an evaluation scale. Although FIG. 18 shows only the pixels in the vertical direction, the pixels in the horizontal direction are also targets for calculating the sum of absolute differences between pixels as a first evaluation scale.

Finally, in step S1103, an interpolation method that minimizes the evaluation scale is selected from among the first evaluation scales obtained by using the motion vector and the still motion vector in the above processing, and the inter-field interpolation pixel is determined. Generate.

Next, an operation example of the inter-field interpolation pixel generator 604 will be described with reference to FIG. In FIG.
The operations in steps S1001 to S1007 are the same as those described with reference to FIG. In step S1201, it is determined whether a valid motion vector (forward, backward, and bidirectional) includes a still motion vector. Step S120
If the result of the determination in step 1 is "No", step S
Proceed to 1203. If “Yes”, the process proceeds to step S1202.

In step S1202, when the second evaluation scale is obtained using the still motion vector, the evaluation scale may be obtained in the same manner as described with reference to FIGS. 14, 15, and 16. Finally, step S1203
Then, among the second evaluation scales obtained by using the motion vector and the still motion vector in the above processing, an interpolation method that minimizes the evaluation scale is selected, and an inter-field interpolation pixel is generated.

As described above, the generated inter-field interpolation pixels are output to the weighting factor determiner 206 and the sequential scanning image generator 607. Next, the operation of the intra-field interpolation pixel generator 605 will be described. The intra-field interpolation pixel generator 605 generates intra-field interpolation pixels using the pixels in the field to be sequentially converted.

A first operation example of the intra-field interpolation pixel generator 605 will be described with reference to FIG. FIG. 19 is a schematic diagram showing a state of pixels on a field to be subjected to progressive scan conversion in a decoded image (interlaced image). FIG.
When an intra-field interpolation pixel is generated for the block c, the pixel is generated by an average value of upper and lower pixels. For example, the interpolation pixel h is generated as an average value of the pixel i and the pixel j.

A second operation example of the intra-field interpolation pixel generator 605 will be described with reference to FIGS. FIG.
Is a diagram illustrating pixels on a field to be sequentially scanned and converted, and illustrates a case where an interpolation pixel h is generated. FIG. 21 is a flowchart showing the operation of the intra-field interpolation pixel generator 605.

First, in step S2101, it is determined whether or not the absolute value of the difference between pixels in the vertical direction (pixel i and pixel q) is greater than or equal to a threshold value. If the result of the determination in step S2101 is “Yes”, the flow proceeds to step S2102. If “No”, step S21
Proceed to 06.

In step S2102, the direction of the edge passing through the pixel h is detected using the peripheral pixels at the interpolation pixel position (pixel h). This corresponds to 2 in the edge direction passing through the pixel h.
This can be achieved by selecting the direction in which the absolute value of the pixel difference is the smallest. That is, 2 in 5 directions indicated by arrows in FIG.
The difference absolute value between pixels is calculated. Here, it is assumed that among the absolute values of the differences in the five directions, the absolute value of the difference between the pixel n and the pixel o is minimized. In that case, a temporary interpolated pixel is generated by the average value of the pixel n and the pixel o.

Next, step S2103, step S2
By the process of 104, the strength of correlation between pixels existing in the direction of the edge is obtained as the reliability of the edge. Step S
In 2103, it is determined whether the ratio of the difference absolute value in the edge direction (pixel n and pixel o) to the difference absolute value in the vertical direction (pixel i and pixel q) is smaller than a threshold value. If the determination result of step S2103 is “Yes”, the process proceeds to step S2104. If “No”, the process proceeds to step S2106.

In step S2104, step S21
It is determined whether the value of the temporary interpolated pixel generated in 02 is a value between pixel values in the vertical direction (pixel i and pixel q).
If the result of the determination in step S2104 is “Yes”, the flow proceeds to step S2105. If “No”, the process proceeds to step S2106. Thus, in step S2105, the temporary interpolated pixel generated in step S2102 is adopted as the interpolated pixel h. On the other hand, in step S2106, an interpolation pixel h is generated using the upper and lower pixels (pixel i and pixel q) in the vertical direction.

As described above, the intra-field interpolation pixel generator 605 performs the above operation for all the interpolation positions in the block to generate the intra-field interpolation pixels.
The generated intra-field interpolation pixels are output to the inter-field interpolation pixel generator 604, the weight coefficient determiner 606, and the progressive scan image generator 607.

Next, the operation of the weight coefficient determiner 606 will be described. The weight coefficient determiner 606 determines a weight coefficient used when the progressively scanned image generator 607 weights and adds the inter-field interpolation pixel and the intra-field interpolation pixel. Here, the determination of the weight coefficient is performed in block units.

FIG. 22 is a flowchart showing the operation of the weight coefficient determiner 606. First, in step S2201, an evaluation scale is calculated using the inter-field interpolation pixel input from the inter-field interpolation pixel generator 604 and the intra-field interpolation pixel input from the intra-field interpolation pixel generator 605. Here, for example, as the evaluation scale, there is a sum of the absolute difference values of the corresponding pixels, a maximum value of the absolute difference values, and a ratio of the average value of the pixels in the block. Here, a case will be described in which the sum of absolute differences of the corresponding pixels is used as the evaluation scale.

Steps S2202 and S2203
Then, the evaluation scale is compared with the threshold. Here, if the evaluation scale (the sum of absolute differences between the inter-field interpolation pixel and the intra-field interpolation pixel) is smaller than the threshold value TH1, the weight coefficient w is set to 1.0 (step S220).
6). Further, when the evaluation scale is equal to or larger than the threshold value TH1 and T
If it is smaller than H2, the weight coefficient w is set to 0.5 (step S2205). When the evaluation scale is the threshold value TH2
If so, the weight coefficient w is set to 0.0 (step S
2204). Here, the relationship between TH1 and TH2 is TH1 <TH2. Then, the determined weighting factors are output to the sequentially scanned image generator 607. Here, the weight coefficient w indicates a weight for inter-field interpolation.

The progressive scan image generator 607 receives the inter-field interpolation pixel block from the inter-field interpolation pixel generator 604, the intra-field interpolation pixel block from the intra-field interpolation pixel generator 605, and the weight coefficient w from the weight coefficient determiner 606. Enter Then, the sum of a value obtained by multiplying each pixel of the inter-field interpolation pixels by a weight coefficient w and a value obtained by multiplying each pixel of the intra-field interpolation pixels by a value of 1.0-w (weight coefficient) is obtained. Let it be an interpolation pixel. Then, a decoded image (interlaced image) input from the image memory 507 is interpolated using these interpolation pixels, and a progressive image is generated and output.

As described above, according to the third embodiment, in the case where a code string obtained by encoding an interlaced image using motion compensation is decoded and the decoded interlaced image is sequentially scanned and converted into a progressive image, A motion vector at the time of motion compensation obtained at the time of decoding is converted into a size in units of one field by a motion vector converter 602, and the validity of the motion vector converted in units of one field is determined by a motion vector determiner 603. Then, the inter-field interpolation pixel generator 604 obtains pixels from the reference field using the motion vector converted into the size of one field unit and the result of determining the validity of the motion vector, and obtains a pixel from the reference scan conversion target field. Inter-field interpolation pixels are generated. Therefore, a more accurate interpolated pixel can be obtained from the nearest field based on the motion vector converted in units of one field, and the image quality of the image after the sequential scan conversion can be improved.

In the third embodiment, the progressively scanned image generator 607 determines the final interpolated pixel by using the inter-field interpolated pixel and the intra-field interpolated pixel by the weight coefficient determining unit 6.
The weighted average determined by the weight coefficient determined in step 06 is generated. Therefore, the possibility of generating an incorrect interpolation pixel can be minimized.

In the third embodiment, when the inter-field interpolation pixel generator 604 generates a plurality of valid motion vectors when generating inter-field interpolation pixels,
An evaluation scale for selecting which motion vector should be used to generate an inter-field interpolation pixel is calculated, and an optimal motion vector for generating an inter-field interpolation pixel is selected using the obtained evaluation scale. Therefore, a motion vector more suitable for progressive scan conversion can be used, and the image quality of the image after progressive scan conversion can be further improved.

In the third embodiment, when the intra-field interpolation pixel generator 605 generates the intra-field interpolation pixel, the direction of the edge at the interpolation position where the interpolation pixel is generated is determined by determining the peripheral pixels of the interpolation position. If the reliability of the detected edge is equal to or more than a predetermined value, an interpolated pixel is generated using pixels existing in the direction of the edge, and the reliability of the edge is determined. When the value is smaller than the value, the interpolation pixel is generated using the pixels existing in the vertical direction of the interpolation position. Therefore, it is possible to improve the image quality of a portion, such as a diagonal line, in the image after the sequential scan conversion.

In the third embodiment, a code string of an interlaced image (MPEG video) encoded by the MPEG system is input. However, the present invention is not limited to this, and the present invention is not limited to this. As long as the method is a method of predictive encoding using compensation, the input code string may be encoded by another method.

Further, in the third embodiment, the case has been described where the motion vector converter 602 converts a motion vector into a motion vector in units of one field, but this may be performed in units of one frame. Third Embodiment
Then, the motion vector converter 602 sets the motion vector to 1
Although the case where the motion vector is converted into the motion vector in the field unit has been described, the amount of motion in the vertical direction may be limited to an even pixel amount. In this case, the number of motion vectors determined to be invalid by the motion vector determiner 603 is reduced, and higher image quality can be achieved.

In the third embodiment, the validity of a motion vector is determined using the motion vector determiner 603. However, all motion vectors may be validated without using the motion vector determiner 603. Third Embodiment
Then, the motion vector determiner 603 determines whether the magnitude of the motion vector converted by the motion vector converter 602 is equal to or smaller than a threshold value and determines whether the vertical motion amount is an even number. As described above, whether the magnitude of the motion vector is equal to or smaller than the threshold value may be determined by the motion vector converter 602 using the motion vector value before the conversion.

Also, in the third embodiment, when the inter-field interpolation pixel generator 604 generates the inter-field interpolation pixel, the motion vector evaluation unit 603 determines that the motion vector evaluation scale is effective. A first evaluation scale calculation method using a sum of pixel difference absolute values of a pixel of the reference field pointed to by the determined effective motion vector and a pixel of the reference field pointed to by the motion vector in the opposite direction to the effective motion vector; , Motion vector determiner 6
03, a pixel in a reference field indicated by an effective motion vector determined to be effective, and an intra-field interpolation pixel generator 605.
And the second evaluation scale calculation method using the sum of absolute differences between pixels with the interpolated pixel generated by the above as an evaluation scale. However, other parameters, such as the maximum of the absolute difference value of the corresponding pixel, are used as the evaluation scale. The value, the ratio of the average value of the pixels in the block, may be used.

In the third embodiment, when the inter-field interpolation pixel generator 604 generates inter-field interpolation pixels, a unit (block) of 8 horizontal pixels and 4 vertical pixels is used.
, But the size of the block is not limited to this. In the third embodiment, the case has been described where the inter-field interpolation pixel generator 604 calculates the evaluation scale for one direction, the opposite direction, and the bidirectional direction of the motion vector when obtaining the evaluation scale of the motion vector. , May not be all calculated but only a part.

In the third embodiment, the case has been described where the intra-field interpolation pixel generator 605 uses the difference absolute value of two pixels in the edge direction when determining the edge direction. The difference value may be calculated by using this. For example, in FIG. 20, when detecting the edge in the direction connecting the pixel m and the pixel p, the pixels i and o, the pixel m
And p, and the sum of the absolute differences of the pixels n and q. By using the peripheral pixels in the calculation of the difference value in this way, it is possible to further prevent erroneous detection of an edge in an image on which noise is present or an image with a very fine picture.

In the third embodiment, the case where the intra-field interpolation pixel generator 605 generates an interpolation pixel from two pixels, that is, the case where the number of taps of the filter when generating the interpolation pixel is two, has been described. The number of taps may be another value. Further, in the third embodiment, the case where the intra-field interpolation pixel generator 605 finds the edge direction from among the five directions has been described, but the number of directions may be another value.

Further, in the third embodiment, the weight coefficient w determined by the weight coefficient determiner 606 has three stages (1, 0.
Although the case of (5, 0) has been described, this may be performed in any number of stages. In the third embodiment, the case has been described where the weighting factor determiner 606 uses the sum of absolute differences between the inter-frame interpolated pixels and the inter-field interpolated pixels as an evaluation scale, but this is obtained by another method. It may be an evaluation scale, the ratio of the maximum difference absolute value or the pixel average value,
And so on. Also, a combination of these may be used as an evaluation scale.

In the third embodiment, the interpolated pixels generated from the inter-field interpolated pixels generated by the inter-field interpolated pixel generator 604 and the intra-field interpolated pixels generated by the intra-field interpolated pixel generator 605 are described. The progressive image is generated by interpolating the decoded image using the interpolation image. The decoded image is interpolated using only the intra-field interpolation pixels generated by the intra-field interpolation pixel generator 605,
A progressive image may be generated. In this case, only the intra-field interpolation pixel generator 605 and the progressive scan image generator 607 operate in the progressive scan converter 508.

Embodiment 4 Hereinafter, Embodiment 4 of the present invention will be described. In the fourth embodiment, a case will be described in which a macroblock to be subjected to progressive scan conversion to be subjected to progressive scan conversion processing is a macroblock that has been intra-coded. In addition, as a macroblock that is intra-coded, in a P picture and a B picture, a case in which a certain macroblock in a frame is intra-coded, and a case in which the frame itself is intra-coded. The case is considered. In general, a macroblock that is intra-coded does not have a motion vector, so that the motion vector cannot be used for progressive scan conversion. Therefore, the fourth embodiment
In P-pictures and B-pictures, if a macroblock to be subjected to progressive scan conversion in a frame is intra-coded, the motion vector of a peripheral macroblock or a preceding or succeeding frame, and the frame itself is an intra-coding. In the case of a structured intra frame, the motion vectors of the preceding and following frames are used for progressive scan conversion. Embodiment 4 is different from Embodiment 3 in that
This is the operation of the motion vector converter 602. The operation of motion vector converter 602 according to Embodiment 4 will be described with reference to FIG.

FIG. 23 is a schematic diagram showing a frame to be subjected to progressive scan conversion in which an intra-coded macroblock exists, and frames before and after the frame. Motion vector converter 6
In step 02, the macro block type of the macro block to be subjected to the sequential scan conversion is read from the parameter memory 601. If it is found from the macroblock type that the macroblock to be progressively scanned is intra-coded, the motion vector of another macroblock is obtained. Here, the other macroblock refers to a peripheral macroblock in the same frame of the sequential scan conversion target macroblock, or a macroblock at the same position in the preceding and succeeding frames. In FIG. 23, macro blocks T, U, V,
W indicates a peripheral macroblock in the same frame, X indicates a macroblock at the same position in the previous frame, and Y indicates a macroblock at the same position in the subsequent frame. Then, when performing sequential scan conversion on the macroblock to be subjected to progressive scan conversion, a motion vector of one macroblock among these macroblocks or a motion vector of a plurality of macroblocks may be used. If the frame to be subjected to progressive scan conversion is intra-coded, the motion vector of the macroblock at the same position in the preceding and succeeding frames with respect to the macroblock to be subjected to progressive scan conversion is used.

The motion vector converter 602 converts the magnitude of the motion vector in units of one field, and outputs the result to the motion vector determiner 603. Since the conversion processing of the motion vector converter 602 and the processing after the motion vector determination unit 603 are the same as those in the third embodiment,
Description is omitted.

As described above, in the fourth embodiment, when a code string obtained by encoding an interlaced image using motion compensation is decoded, and the decoded interlaced image is sequentially scanned and converted into a progressive image, If the macroblock to be subjected to progressive scan conversion is intra-coded, the motion vector of the peripheral macroblock in the same frame or the macroblock at the same position in the preceding and succeeding frames is reduced to the size of one field unit. The validity of the motion vector converted at 602 and converted to the size of one field unit is determined by the motion vector determiner 603, and the motion converted to the size of one field unit is calculated by the inter-field interpolation pixel generator 604. A pixel is obtained from the reference field using the vector and the determination result of the validity of the motion vector, and the sequential scanning change is performed. And to generate a field between the interpolation pixel with respect to the target field. Therefore, even if the macroblock to be subjected to progressive scan conversion is an intra-coded macroblock, it is possible to generate inter-frame interpolation pixels with high accuracy without performing motion detection.

(Embodiment 5) Embodiment 5 of the present invention
Will be described with reference to FIG. In the fifth embodiment, M
The progressive scan conversion in the case where the MPEG video code string input to the PEG video decoder 501 is a code string read from a recording medium in fast-forward or fast-reverse mode will be described. Generally, when an MPEG stream is read from a recording medium in a fast-forward or fast-reverse mode, only an I-picture is read, or an I-picture and a P-picture are read.
Only pictures are read. That is, when only the I picture is read, the input code string becomes a code string including only the data of the I picture,
When only a picture is read, the input code string is a code string including only the data of the I picture and the P picture. Therefore, when the MPEG video code string input to the MPEG video decoder 501 is a code string read from the recording medium in the fast-forward or fast-reverse mode, the progressive scan converter 508 uses the motion vector to It is not possible to acquire pixels from a field and generate inter-field interpolation pixels to perform sequential scan conversion. Therefore, in Embodiment 5, the progressive scan converter 508 says that the MPEG video code string input by the MPEG video decoder 501 is a code string read from the recording medium in the fast-forward or fast-reverse mode. When the reproduction method information is input, a progressive image is generated using only the intra-field interpolation pixels. That is, the motion vector converter 602 does not read the motion vector from the parameter memory 601. Therefore, the motion vector determiner 603 and the inter-field interpolation pixel generator 604 do not perform any processing. Further, the weight coefficient determiner 606 outputs the weight coefficient 0 (that is, the in-field interpolation pixel is used as the interpolation pixel as it is) to the progressive scanning image generator 607. Then, the progressive scanning image generator 607 generates a progressive image by using the intra-field interpolation pixels output from the intra-field interpolation pixel generator 605 as interpolation pixels as they are.

As described above, in the fifth embodiment, the MP
If the MPEG video code string input to the EG video decoder 501 is a code string read from the recording medium in the fast forward or fast reverse mode, the progressive scan image generator 60
No. 7 generates an interpolated pixel from only the intra-field interpolated pixels generated by the intra-field interpolated pixel generator 605, and generates a progressive image using the interpolated pixels. Therefore, even when the input code string is a code string read out from the recording medium in the fast forward or fast reverse mode, the sequential scan conversion process can be performed without providing a new component and increasing the circuit scale. It can be carried out. In addition,
In the case of the fast-forward or fast-return mode, the reproduced image changes at a high speed. Therefore, even if the interpolation pixels are generated only by the interpolation pixels in the field, the deterioration of the vertical resolution is not noticeable.

[0150]

As described above, according to the progressive scan conversion method and the progressive scan conversion apparatus of the present invention, when converting an interlaced image into a progressive image, the field to be subjected to the progressive scan conversion and the pixels in the preceding and succeeding fields are converted. V for
An interpolation pixel is generated by applying a T filter. At that time, the sum of the absolute differences between pixels between the frame including the field to be subjected to the progressive scan conversion and the field to be subjected to the progressive scan conversion and the field or frame immediately before or after the frame is obtained, and based on the value of the sum of the absolute values. , VT filter coefficients. This coefficient is set so that the larger the sum of absolute differences is, the more the moving image is determined to be a moving image and the contribution (gain) from the adjacent field is reduced. Therefore, in the case of a still image, a high-resolution progressive image can be obtained by the conventional VT filter, and in the case of a moving image, the image quality of a portion where the image quality has deteriorated in the conventional VT filter can be greatly improved. Can be. In addition, these operations are 1
Since it can be realized with one filter, cost can be reduced.

Further, when the progressive scan conversion method and the progressive scan conversion device of the present invention convert an interlaced image into a progressive image, the progressive scan conversion method and the progressive scan conversion device perform the progressive scan conversion on a field to be progressively converted and pixels in fields before and after the field. An interpolation pixel is generated by applying a VT filter. At this time, an absolute difference between pixels in adjacent fields near the interpolation position is calculated, and VT is calculated based on the absolute value.
Determine the coefficients of the filter. The coefficient is determined such that the larger the sum of absolute differences is, the more the moving image is determined, and the smaller the contribution from the adjacent field is. Therefore, in the case of a still image, a high-resolution progressive image can be obtained by the conventional VT filter, and when there is an object moving on the screen, the object is replaced with the conventional VT filter. It is possible to prevent image quality degradation that has occurred in the filter. Further, since these operations can be realized by one filter, cost can be reduced.

Further, according to the progressive scan conversion method and the progressive scan converter of the present invention, a code sequence obtained by encoding an interlaced image using motion compensation is decoded, and the decoded image of the interlaced image is decoded together with the decoded image. A motion vector is obtained, the motion vector is converted to a size in units of one field, the validity of the motion vector converted to a size in units of one field is determined, and the motion vector and the determination result of the motion vector are determined. And obtains pixels from the reference field, generates inter-field interpolation pixels for the field to be subjected to progressive scan conversion, generates intra-field interpolation pixels using pixels in the field to be subjected to progressive scan conversion, and inter- and inter-field interpolation pixels. Interpolated pixels are generated by weighting and averaging the pixels with weight coefficients, and decoding is performed using the interpolated pixels. Since the progressive scan image is generated by interpolating the image, a more accurate interpolated pixel can be obtained from the nearest field based on the motion vector converted in units of one field. Higher image quality can be achieved. In addition, the final interpolation pixel is generated by a weighted average of the inter-field interpolation pixel and the intra-field interpolation pixel, and this weight is determined by the difference between the inter-field interpolation pixel and the intra-field interpolation pixel. The possibility of generating an interpolated pixel can be minimized.

Further, according to the progressive scan conversion method and the progressive scan conversion apparatus of the present invention, if there are a plurality of effective motion vectors that can be used for generating inter-field interpolation pixels, Calculate a rating scale to select whether to generate interpolated pixels,
By selecting the optimal motion vector for the field interpolation processing using the obtained evaluation scale, a motion vector more suitable for the progressive scan conversion can be used, and the image quality of the image after the progressive scan conversion can be further improved. it can.

Further, according to the progressive scan conversion method and the progressive scan conversion apparatus of the present invention, when generating an intra-field interpolation pixel, the edge direction at the interpolation position where the interpolation pixel is generated is determined by using the peripheral pixels of the interpolation position. The reliability of the detected edge is determined.If the reliability of the edge is equal to or more than a predetermined value, an interpolation pixel is generated using pixels existing in the direction of the edge. If the value is less than 1, since the interpolation pixel is generated using the pixels existing in the vertical direction of the interpolation position, it is possible to improve the image quality of a portion, such as a diagonal line, in the image after the sequential scan conversion, particularly. it can.

Further, according to the progressive scan conversion method and the progressive scan converter of the present invention, a code sequence obtained by encoding an interlaced image using motion compensation is decoded, and the decoded image of the interlaced image is decoded together with the decoded image. When acquiring a motion vector, if the macroblock to be sequentially scanned and converted is intra-coded, the motion vector of a peripheral macroblock in the same frame or a macroblock at the same position in the previous and next frames is calculated in units of one field. The motion vector is converted into a size, the validity of the motion vector converted into the size of one field unit is determined, and the pixels obtained from the reference field using the motion vector and the determination result of the motion vector are used to sequentially scan the field to be converted. Generate interpolated pixels between fields, and perform field conversion using pixels in the field to be progressively converted. An interpolated pixel is generated, an interpolated pixel is generated by weighting and averaging the inter-field interpolated pixel and the intra-field interpolated pixel with a weight coefficient, and a progressive image is generated by interpolating a decoded image using the interpolated pixel. In this way, even in the case of an intra-coded image, it is possible to generate inter-frame interpolation pixels with high accuracy without the need for motion detection, and to improve the image quality of the image after progressive scan conversion. Can be planned.

Further, according to the progressive scan conversion method and the progressive scan conversion apparatus of the present invention, when the input code string is a code string read out from the recording medium in the fast forward or fast reverse mode, the Since only the interpolated pixels are generated and the progressive image is generated using only the generated interpolated pixels in the field, fast forward or rewind from the recording medium without providing a new component and increasing the circuit scale. The code string read in the mode can be sequentially scan-converted.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a progressive scan conversion device according to Embodiment 1 of the present invention.

FIG. 2 is a schematic diagram for explaining an interlaced image.

FIG. 3 is a VT filter device 10 according to the embodiment of the present invention.
FIG. 4 is a schematic diagram for explaining an example of a frequency characteristic of No. 2;

FIG. 4 is a block diagram showing a configuration of a progressive scan conversion device according to Embodiment 2 of the present invention.

FIG. 5 is a block diagram showing a configuration of a progressive scan conversion device according to Embodiment 3 of the present invention.

FIG. 6 shows a progressive scan converter 50 according to an embodiment of the present invention.
8 is a block diagram of FIG.

FIG. 7 is a schematic diagram showing a state of an interlaced image.

FIG. 8 is a flowchart for explaining a determination method of a motion vector determiner 603 according to the embodiment of the present invention.

FIG. 9 is a flowchart illustrating an operation example of the inter-field interpolation pixel generator 604 according to the embodiment of the present invention.

FIG. 10 is a flowchart illustrating an operation example of the inter-field interpolation pixel generator 604 according to the embodiment of the present invention.

FIG. 11 is a flowchart illustrating an operation example of the inter-field interpolation pixel generator 604 according to the embodiment of the present invention.

FIG. 12 is a flowchart illustrating an operation example of the inter-field interpolation pixel generator 604 according to the embodiment of the present invention.

FIG. 13 is a schematic diagram showing a state of an interlaced image.

FIG. 14 is a schematic diagram showing a state of an interlaced image.

FIG. 15 is a schematic diagram showing a state of an interlaced image.

FIG. 16 is a schematic diagram showing a state of an interlaced image.

FIG. 17 is a schematic diagram showing a state of an interlaced image.

FIG. 18 is a schematic diagram illustrating a state of an interlaced image.

FIG. 19 is a schematic diagram showing a state of a field pixel to be subjected to progressive scan conversion.

FIG. 20 is a schematic diagram illustrating a state of a field pixel to be sequentially scanned and converted;

FIG. 21 is a flowchart illustrating an operation example of the intra-field interpolation pixel generator 605 according to the embodiment of the present invention.

FIG. 22 is a diagram illustrating a weight coefficient determiner 6 according to an embodiment of the present invention.
It is a flowchart figure which shows operation | movement of 06.

FIG. 23 is a schematic diagram showing a frame to be subjected to progressive scan conversion including an intra-coded macroblock, and frames before and after the frame.

FIG. 24 is a block diagram showing a configuration of a progressive scan converter 508 according to Embodiment 5 of the present invention.

FIG. 25 is a schematic diagram of an interlaced image for explaining a conventional progressive scan conversion method.

[Explanation of symbols]

 Reference Signs List 101 Frame memory 102 VT filter 103 Filter coefficient setting unit 104, 404 Difference calculator 105 Double speed converter 501 MPEG video decoder 502 Variable length decoder 503 Inverse quantizer 504 Inverse DCT unit 505 Adder 506 Motion compensator 507 Image memory 508 Progressive scan converter 601 Parameter memory 602 Motion vector converter 603 Motion vector determiner 604 Inter-field interpolation pixel generator 605 In-field interpolation pixel generator 606 Weight coefficient determiner 607 Progressive scan image generator

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5C059 KK19 LA01 LA06 LB13 MA03 MA23 NN01 NN21 PP04 RC17 TA08 TA62 TA68 TC12 UA05 UA11 5C063 AA01 AB03 BA04 BA12 CA05 CA07 CA09 CA23 CA40

Claims (37)

[Claims] 1. A progressive scan conversion method for converting an interlaced scan image into a progressive scan image, comprising: a progressive scan conversion target field to be subjected to progressive scan conversion processing; Generating an interpolated pixel of the field to be subjected to progressive scan conversion by applying a filtering process to pixels of at least one of the three fields, and measuring a motion amount of the field to be subjected to progressive scan conversion And a step of changing a characteristic of the filtering process based on the amount of motion. 2. The progressive scan conversion method according to claim 1, wherein the filter used for the filtering in the step of generating the interpolated pixels includes a low frequency component in a vertical direction for the field to be progressively converted. And for the fields before and after the field to be subjected to the progressive scan conversion,
A progressive scan conversion method having a characteristic of extracting a high frequency component in a vertical direction. 3. The progressive scan conversion method according to claim 1, wherein the step of generating the interpolated pixel includes performing a filtering process on a peripheral pixel in a vertical-time plane of an interpolation position in the field to be subjected to the progressive scan conversion. And a progressive scan conversion method. 4. The progressive scan conversion method according to claim 1, wherein the step of measuring the amount of motion comprises: performing the progressive scan conversion target field or the frame including the progressive scan conversion target field; A progressive scan conversion method, wherein the amount of motion is obtained from a difference value between. 5. The progressive scan conversion method according to claim 1, wherein the step of measuring the amount of motion is performed based on a difference value between pixels used when performing the filtering in the step of generating the interpolation pixel. A progressive scan conversion method characterized by determining an amount. 6. The progressive scan conversion method according to claim 5, wherein the step of measuring the amount of motion includes the progressive scan conversion among pixels used when applying the filter in the step of generating the interpolation pixel. A progressive scan conversion method characterized in that a motion amount is obtained from a difference value between pixels belonging to fields before and after a target field. 7. The progressive scan conversion method according to claim 1, wherein the step of changing the characteristics of the filter processing includes changing the characteristics of the filter processing before and after the field to be subjected to the progressive scan conversion as the amount of motion increases. A progressive scan conversion method, wherein the gain of a component from a field is changed to be small. 8. The progressive scan conversion method according to claim 1, wherein the step of changing the characteristics of the filter processing includes changing the characteristics of the filter processing when the amount of motion is large. A progressive scan conversion method characterized by changing the gain of the component from the image data to zero. 9. A progressive scan conversion device for converting an interlaced scan image into a progressive scan image, comprising: a frame memory for accumulating the interlaced scan image;
A progressive scan conversion target field to be subjected to progressive scan conversion processing, and either or both of a preceding field and a subsequent field of the progressive scan conversion target field are input, and at least one of the input fields is input.
A filter that generates an interpolation pixel of the field to be subjected to the progressive scan conversion by performing a filtering process on the pixels of the field; a difference calculator that measures the amount of motion of the field to be subjected to the progressive scan conversion; And a filter coefficient setting device for changing a characteristic of the filter device based on the amount of motion obtained. 10. A progressive scan conversion device for converting an interlaced scan image into a progressive scan image, comprising: a frame memory for accumulating the interlaced scan image;
A progressive scan conversion target field to be subjected to progressive scan conversion processing, and either or both of a preceding field and a subsequent field of the progressive scan conversion target field are input, and at least one of the input fields is input.
A filter that generates an interpolated pixel of the field to be subjected to the progressive scan conversion by performing a filter process on the pixels of the field; and a frame to be subjected to the progressive scan conversion or the field to be subjected to the progressive scan conversion in the interlaced scan image from the frame memory. A frame including a field and a field or a frame adjacent to the frame including the field to be subjected to the progressive scan conversion or the field to be subjected to the progressive scan conversion are input, and the difference between them is calculated to obtain the motion amount of the field to be subjected to the progressive scan conversion A filter coefficient setting device that changes the filter characteristic of the filter device based on the motion amount measured by the difference calculation device; an interlaced scan image; and an interpolation generated by the filter device. Generates progressively scanned images by combining with pixels Sequential scanning conversion apparatus, characterized in that it comprises a speed transducer, the. 11. A decoding process is performed for each field or frame for a code string obtained by encoding a skip scanning image composed of a plurality of fields using motion compensation, and a skip scanning image obtained by the decoding process is obtained. In the progressive scan conversion method of converting to a progressive scan image, a decoded image is obtained by performing a decoding process on the interlaced scan image, and at the time of the motion compensation, a motion vector indicating a predetermined reference field for a target field is calculated. Converting the motion vector having a size corresponding to a time interval from the target field to the predetermined reference field in each of the fields into a motion vector having a size corresponding to a fixed unit time interval. A motion vector conversion step and a progressive scan conversion target to be subjected to progressive scan conversion processing An inter-field interpolation pixel generating step of obtaining pixels based on a motion vector converted by the motion vector conversion step from a reference field which is a field before and after the field, and generating a first interpolation pixel for the progressive scan conversion target field And an intra-field interpolation pixel generation step of generating a second interpolation pixel using a pixel in the progressive scan conversion target field; and a weighting ratio between the first interpolation pixel and the second interpolation pixel. A weighting factor determining step of determining a weighting factor to be indicated, the first interpolation pixel and the second interpolation pixel are weighted and averaged by the weighting factor to generate a third interpolation pixel, the third interpolation pixel A progressively scanned image generating step of generating a progressively scanned image by interpolating the decoded image using an interpolation pixel. A progressive scan conversion method characterized by the above-mentioned. 12. A decoding process is performed on a code string obtained by encoding a skip scanning image composed of a plurality of fields using motion compensation for each field or frame, and the skip scanning image obtained by the decoding process is obtained. In the progressive scan conversion method of converting to a progressive scan image, a decoded image is obtained by performing a decoding process on the interlaced scan image, and at the time of the motion compensation, a motion vector indicating a predetermined reference field for a target field is calculated. Converting the motion vector having a size corresponding to a time interval from the target field to the predetermined reference field in each of the fields into a motion vector having a size corresponding to a fixed unit time interval. A motion vector conversion step; A motion vector determining step of determining the validity of the motion vector; and a motion vector and the motion converted by the motion vector converting step from a reference field which is a field before and after a field to be subjected to progressive scan conversion. A pixel is obtained based on the determination result in the vector determination step, an inter-field interpolation pixel generation step of generating a first interpolation pixel for the progressive scan conversion target field, and using a pixel in the progressive scan conversion target field An intra-field interpolation pixel generation step of generating a second interpolation pixel, a weight coefficient determination step of determining a weight coefficient indicating a weight ratio between the first interpolation pixel and the second interpolation pixel, A weighted average of one interpolated pixel and the second interpolated pixel by the weight coefficient Generating a third interpolated pixel, and interpolating the decoded image using the third interpolated pixel to generate a progressively scanned image. Scan conversion method. 13. The sequential scan conversion method according to claim 11, wherein the time interval of a fixed unit in the motion vector conversion step is a time interval corresponding to one field. Scan conversion method. 14. The progressive scan conversion method according to claim 11, wherein the processing of the inter-field interpolation pixel generation step, the weight coefficient determination step, and the progressive scan image generation step are performed during the motion compensation. A progressive scan conversion method, wherein the motion vector is performed in a unit smaller than the associated image unit. 15. The progressive scan conversion method according to claim 11, wherein the code string is a code string encoded by an MPEG system. 16. The progressive scan conversion method according to claim 11, wherein, when the distance between one line of the frame structure is one pixel, the vertical component of the motion vector is determined. A progressive scan conversion method, wherein conversion is performed so as to be an even value. 17. The progressive scan conversion method according to claim 12, wherein the motion vector determination step determines that the motion vector is valid if the magnitude of the motion vector converted in the motion vector conversion step is equal to or smaller than a predetermined value. And a progressive scan conversion method. 18. The progressive scan conversion method according to claim 12, wherein, when the distance between one line of the frame structure is one pixel, the motion vector determination step is performed on the motion vector converted by the motion vector conversion step. A progressive scan conversion method, wherein a motion vector whose vertical component is an even value is determined to be valid. 19. The progressive scan conversion method according to claim 11, wherein the inter-field interpolation pixel generation step is optimal for generating the first interpolation pixel using the motion vector converted by the motion vector conversion step. A method for calculating an evaluation scale for selecting a suitable motion vector, and generating the first interpolation pixel using a motion vector having the best evaluation scale. 20. The progressive scan conversion method according to claim 11, wherein the inter-field interpolation pixel generation step includes: a motion vector converted by the motion vector conversion step;
Using a motion vector in the opposite direction to the motion vector, an evaluation scale for selecting a motion vector that is optimal for generating the first interpolation pixel is calculated. In the step of generating the first interpolation pixel, the motion vector in the reverse direction, the direction is opposite to the motion vector converted by the motion vector conversion step,
And a reference field indicated by the motion vector is a motion vector indicating a reference field whose context is opposite to that of a target field. 21. The progressive scan conversion method according to claim 12, wherein the inter-field interpolation pixel generation step is determined to be valid in the motion vector determination step among the motion vectors converted by the motion vector conversion step. Calculating an evaluation scale for selecting a motion vector optimal for the generation of the first interpolation pixel using the motion vector,
A progressive scan conversion method, wherein the first interpolation pixel is generated using a motion vector having the best evaluation scale. 22. The progressive scan conversion method according to claim 12, wherein the inter-field interpolation pixel generation step is determined to be valid in the motion vector determination step among the motion vectors converted by the motion vector conversion step. Using the calculated effective motion vector and the motion vector in the opposite direction to the motion vector, calculate an evaluation scale for selecting a motion vector that is optimal for generating the first interpolation pixel. Generating the first interpolation pixel using a motion vector becomes, the reverse motion vector, the direction is opposite to the effective motion vector, and the reference field indicated by the effective motion vector Is a motion vector pointing to a reference field whose context is opposite to the target field.査 conversion method. 23. The progressive scan conversion method according to claim 19, wherein the inter-field interpolation pixel generation step includes: a motion vector converted by the motion vector conversion step;
An evaluation scale for selecting a motion vector optimal for generating the first interpolation pixel is calculated using the motion vector having the motion of 0, and the first evaluation vector is calculated using the motion vector having the best evaluation scale. A progressive scan conversion method characterized by generating an interpolation pixel of 24. The progressive scan conversion method according to claim 19, wherein the evaluation scale is: a pixel of a reference field indicated by a motion vector converted by the motion vector conversion step; A progressive scan conversion method characterized in that it is a sum of pixel difference absolute values with two interpolation pixels. 25. The progressive scan conversion method according to claim 23, wherein the evaluation scale is a pixel between a pixel of a reference field indicated by the motion vector converted by the motion vector conversion step and the second interpolation pixel. A progressive scan conversion method characterized by being a sum of absolute differences. 26. The progressive scan conversion method according to claim 20, wherein the evaluation scale is a pixel in a reference field indicated by a motion vector converted in the motion vector conversion step, and the motion in the reverse direction. A progressive scan conversion method characterized by being a sum of pixel difference absolute values with a pixel of a reference field indicated by a vector. 27. The progressive scan conversion method according to claim 20, wherein the inter-field interpolation pixel generation step includes: a motion vector converted by the motion vector conversion step;
An evaluation scale for selecting a motion vector optimal for generating the first interpolation pixel is calculated using the motion vector having the motion of 0, and the first evaluation vector is calculated using the motion vector having the best evaluation scale. Wherein the evaluation scale is a pixel difference between a pixel of the reference field indicated by the motion vector converted by the motion vector conversion step and a pixel of the reference field indicated by the reverse motion vector. A progressive scan conversion method characterized by being a sum of absolute values. 28. A progressive scan conversion method for generating an interpolated pixel using a pixel in each field for an interlaced scan image comprising a plurality of fields and converting the interlaced scan image into a progressive scan image. An edge detection step of detecting, as an edge direction, a direction indicated by a straight line connecting peripheral pixels of the interpolation position, passing through the interpolation position that generates the interpolation position, and determining the strength of the correlation between the pixels existing in the edge direction by the edge. An edge reliability determining step for determining the reliability, and when the reliability of the edge is equal to or more than a predetermined value, an interpolation pixel is generated using a pixel existing in the direction of the edge, and the reliability of the edge is a predetermined value. If less than, an interpolated pixel generation step of generating an interpolated pixel using pixels present in the vertical direction of the interpolation position. Next scan conversion method. 29. The progressive scan conversion method according to claim 11, wherein the intra-field interpolation pixel generation step includes passing an interpolation position for generating the second interpolation pixel, and surrounding pixels of the interpolation position. An edge detecting step of detecting a direction indicated by a straight line connecting the two as an edge direction; an edge reliability determining step of obtaining the strength of correlation between pixels existing in the edge direction as the reliability of the edge; If the reliability is equal to or more than a predetermined value, the second interpolation pixel is generated using a pixel existing in the direction of the edge, and when the reliability of the edge is less than a predetermined value,
An interpolated pixel generating step of generating the second interpolated pixel using pixels existing in the up-down direction of the interpolated position. 30. The progressive scan conversion method according to claim 28, wherein, in the edge reliability determining step, the pixel difference value of a pixel existing in the direction of the edge is a pixel in a vertical direction of the interpolation position. Wherein the reliability of the edge is determined to be greater than or equal to a predetermined value if the inter-pixel difference value is smaller than the inter-pixel difference value. 31. The progressive scan conversion method according to claim 29, wherein, in the edge reliability determining step, the pixel difference value of a pixel existing in the direction of the edge is a pixel in a vertical direction of the interpolation position. Wherein the reliability of the edge is determined to be greater than or equal to a predetermined value if the inter-pixel difference value is smaller than the inter-pixel difference value. 32. The progressive scan conversion method according to claim 28, wherein, in the edge reliability determination step, an interpolation pixel value obtained by using a pixel existing in the edge direction exists above and below the interpolation position. A method for determining that the reliability of the edge is equal to or greater than a predetermined value if the value is between the pixel values of the pixels to be converted. 33. The progressive scan conversion method according to claim 29, wherein in the edge reliability determining step, an interpolation pixel value obtained using a pixel existing in the direction of the edge exists above and below the interpolation position. A method for determining that the reliability of the edge is equal to or greater than a predetermined value if the value is between the pixel values of the pixels to be converted. 34. The progressive scan conversion method according to claim 11, wherein the progressive scan conversion is performed on an intra-coded progressive scan conversion target image area on the progressive scan conversion target field. A peripheral image area located around the target image area,
Or performing sequential scan conversion processing using a motion vector associated with an image area at the same position as the sequential scan conversion target image area in a frame immediately before or immediately after the sequential scan conversion target field. Scan conversion method. 35. The progressive scan conversion method according to claim 11, wherein the code string decoded in the decoding step is recorded on a recording medium and read out in a fast forward or fast reverse mode. Wherein the decoded image is interpolated using only the second interpolated pixels generated in the intra-field interpolated pixel generating step to generate a progressively scanned image. Method. 36. A decoding process is performed for each field or frame for a code string obtained by coding an interlaced scan image including a plurality of fields using motion compensation, and the interlaced scan image obtained by the decoding process is obtained. In a progressive scan conversion device for converting to a progressive scan image, a decoded image is obtained by performing a decoding process on the interlaced scan image, and a motion vector indicating a predetermined reference field for a target field at the time of the motion compensation. A decoder to obtain, an image memory for storing the decoded image, a parameter memory for storing the motion vector, and a time from the target field to the predetermined reference field in each field read from the parameter memory. When a motion vector of a size corresponding to the interval is A motion vector converter for converting the motion vector into a motion vector having a size corresponding to the interval; and a motion converted by the motion vector converter from a reference field which is a field before and after a field to be subjected to the progressive scan conversion processing. A pixel is obtained based on a vector, and an inter-field interpolation pixel generator that generates a first interpolation pixel for the progressive scan conversion target field, and a second interpolation pixel using a pixel in the progressive scan conversion target field An in-field interpolation pixel generator to generate, a weight coefficient determiner that determines a weight coefficient indicating a weight ratio between the first interpolation pixel and the second interpolation pixel, and the first interpolation pixel, A third interpolated pixel is generated by averaging the second interpolated pixel with the weighted coefficient using the weighting coefficient, and using the third interpolated pixel. Deinterlacing device characterized by and a progressive scan image generator for generating a progressive scan image by interpolating the decoded image read from the image memory Te. 37. A decoding process is performed for each field or frame for a code string obtained by encoding a skip scanning image including a plurality of fields using motion compensation, and the skip scanning image obtained by the decoding process is obtained. In a progressive scan conversion device for converting to a progressive scan image, a decoded image is obtained by performing a decoding process on the interlaced scan image, and a motion vector indicating a predetermined reference field for a target field at the time of the motion compensation. A decoder to obtain, an image memory for storing the decoded image, a parameter memory for storing the motion vector, and a time from the target field to the predetermined reference field in each field read from the parameter memory. When a motion vector of a size corresponding to the interval is A motion vector converter that converts the motion vector into a motion vector having a size corresponding to the interval; a motion vector determiner that determines the validity of the motion vector converted by the motion vector converter; and a sequential scan conversion processing target. From a reference field that is a field before and after the scan conversion target field, a pixel is obtained based on the motion vector converted by the motion vector converter and the determination result by the motion vector determiner, and a first An inter-field interpolated pixel generator that generates an interpolated pixel, an intra-field interpolated pixel generator that reads out a pixel in the progressive scan conversion target field and generates a second interpolated pixel, the first interpolated pixel and the interpolated pixel A weight coefficient determiner that determines a weight coefficient indicating a weight ratio between the second interpolation pixel and The first interpolated pixel and the second interpolated pixel are weighted and averaged by the weight coefficient to generate a third interpolated pixel, and the third interpolated pixel is read from the image memory using the third interpolated pixel. A progressive scan conversion device comprising: a progressive scan image generator that generates a progressive scan image by interpolating a decoded image.