CN101365093B - Image processing apparatus and method - Google Patents
Image processing apparatus and method Download PDFInfo
- Publication number
- CN101365093B CN101365093B CN2008102102790A CN200810210279A CN101365093B CN 101365093 B CN101365093 B CN 101365093B CN 2008102102790 A CN2008102102790 A CN 2008102102790A CN 200810210279 A CN200810210279 A CN 200810210279A CN 101365093 B CN101365093 B CN 101365093B
- Authority
- CN
- China
- Prior art keywords
- image
- drop
- down pattern
- confirm
- correlation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
- H04N7/0117—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level involving conversion of the spatial resolution of the incoming video signal
- H04N7/012—Conversion between an interlaced and a progressive signal
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
- H04N7/0112—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level one of the standards corresponding to a cinematograph film standard
- H04N7/0115—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level one of the standards corresponding to a cinematograph film standard with details on the detection of a particular field or frame pattern in the incoming video signal, e.g. 3:2 pull-down pattern
Abstract
An image processing apparatus includes: a correlation value computing section that computes a correlation value between pixels in a noted image and its preceding and following reference images; a break determining section that determines a break in identical images based on the correlation value between pixels; a successive occurrence counting section that cumulatively counts successive occurrences of identical images as a successive occurrence count, if the noted image is not determined as corresponding to a break; a pattern determining section that determines whether or not the successive occurrence count and a set count match, if the noted image is determined as corresponding to a break; a matching counting section that counts up a matching count if it is determined that the successive occurrence count and the set count match; and a pull-down pattern detecting section that detects a predetermined pull-down pattern if the matching count exceeds a predetermined count.
Description
Technical field
The present invention relates to image processing apparatus and method and program.More specifically, the present invention relates to make it possible to detect image processing apparatus and the method and the program of various drop-down patterns (pull-down pattern).
Background technology
The technology that is used to change frame frequency (frame rate) is becoming usual day by day.
As the representation of this technology, such technology has been proposed: the frame frequency (for example seeing the open No.2006-253766 of japanese unexamined patent) that through the drop-down downdraw process of so-called 2-3 the frame-rate conversion of the 24fps that uses in the film (frame/per second) is become the 30ftp that uses in the TV thus.
According to the 2-3 downdraw process; For example; The image of first frame of film is used for first and second of television image; The image of second frame of film is used for the 3rd to the 5th of television image, the image of the 3rd frame of film is used for the 6th and the 7th of television image, and the image of the 4th frame of film is used for the 8th to the tenth of television image.Through repeating this process in order, the frame frequency of 24fps is converted into 30fps.In the case, odd field is a top field, and even field is a bottom field.Through odd field and even field continuously, generate a frame of television image.Also there is the drop-down pattern of other type, drop-down such as 2-2.
Summary of the invention
Yet, with the frame frequency of original progressive picture less than 24fps (drop-down like 8-7,6-4 is drop-down or 5-5 drop-down) the corresponding processing of situation need detect through different configurations, this possibly not only increase the complexity of the configuration of device, and causes cost to improve.
Therefore hope to dispose through single process to detect each drop-down pattern, particularly about the input picture of various frame frequencies.
Image processing apparatus comprises according to an embodiment of the invention: the correlation value calculation parts, and the image that is used for calculating continuous input is being paid close attention to image with at said concern image before and the correlation between each of reference picture afterwards; Interrupt to confirm parts, be used for confirming that based on said correlation said concern image is whether corresponding to the interruption of predetermined number of consecutive identical image; Continuous occurrence count parts, if be used for confirming said concern image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count; Pattern is confirmed parts, if be used for confirming said concern image corresponding to interruption, then confirms the setting counting that whether matees the indication continuous occurrence count corresponding with predetermined drop-down pattern in the continuous occurrence count of preceding image up to being right after; The coupling counter block if be used for confirming said concern image corresponding to interruption, and is confirmed to mate said setting counting up to being right after in the continuous occurrence count of preceding image, then counts the number of times that occurs with said predetermined drop-down pattern matched, counts as coupling; And drop-down pattern detection parts; If be used for confirming that said concern image is corresponding to interruption; And confirm to mate said setting counting and said coupling counting greater than preset count up to being right after in the continuous occurrence count of preceding image; Then detect predetermined drop-down pattern, as the drop-down pattern of the image of continuous input corresponding to said setting counting.
This image processing apparatus can be configured to and comprise that further progressive picture generates parts, is used for the drop-down pattern based on the image of the continuous input that is detected by drop-down pattern detection parts, from the image generation progressive picture of continuous input.
Drop-down pattern detection parts can be configured feasible: if confirm to pay close attention to image corresponding to interruption; If then confirm up to the continuous occurrence count match settings counting that is right after at preceding image; And the coupling counting is not more than preset count; But greater than another preset count littler than preset count, then drop-down pattern detection parts suspend the detection to predetermined drop-down pattern.
Interrupt to confirm that parts can be arranged to based on correlation and whether confirm that greater than the comparative result between predetermined threshold or the correlation concern image is whether corresponding to the interruption of predetermined number of consecutive identical image.
Image processing apparatus can be configured to further comprise threshold value management component, is used for managing predetermined threshold so that become littler according to continuous occurrence count.
The correlation value calculation parts can be arranged to calculating, and each is paid close attention to image and is paying close attention to before the image and the correlation between the pixel near the first area the center of reference picture afterwards in the continuous image of importing; And can detect under the state of predetermined drop-down pattern at drop-down pattern detection parts, be used for calculating the image of input continuously greater than the first area and before each is paid close attention to image and is paying close attention to image and the correlation between the pixel near the second area the center of reference picture afterwards.
Image processing method comprises step according to an embodiment of the invention: each that calculate in the continuous image of importing paid close attention to image and paid close attention to before the image and the correlation between the reference picture afterwards; Whether confirm to pay close attention to image corresponding to the interruption in the predetermined number of consecutive identical image based on correlation; If confirm to pay close attention to image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count; If confirm to pay close attention to image, then confirm the setting counting that whether matees the indication continuous occurrence count corresponding in the continuous occurrence count of preceding image up to being right after with predetermined drop-down pattern corresponding to interruption; If confirm to pay close attention to image, and confirm then to count the number of times that occurs with predetermined drop-down pattern matched and count as coupling up to the continuous occurrence count match settings counting that is right after at preceding image corresponding to interruption; And if confirm to pay close attention to image corresponding to interruption; And confirm to count greater than preset count up to the continuous occurrence count match settings counting and the coupling that are right after at preceding image; Then detect corresponding to the predetermined drop-down pattern of setting counting, as the drop-down pattern of the image of continuous input.
Program causes computer to carry out the process that may further comprise the steps according to an embodiment of the invention: each that calculate in the continuous image of importing paid close attention to image and paid close attention to before the image and the correlation between the reference picture afterwards; Whether confirm to pay close attention to image corresponding to the interruption in the predetermined number of consecutive identical image based on correlation; If confirm to pay close attention to image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count; If confirm to pay close attention to image, then confirm the setting counting that whether matees the indication continuous occurrence count corresponding in the continuous occurrence count of preceding image up to being right after with predetermined drop-down pattern corresponding to interruption; If confirm to pay close attention to image, and confirm then to count the number of times that occurs with predetermined drop-down pattern matched and count as coupling up to the continuous occurrence count match settings counting that is right after at preceding image corresponding to interruption; And if confirm to pay close attention to image corresponding to interruption; And confirm to count greater than preset count up to the continuous occurrence count match settings counting and the coupling that are right after at preceding image; Then detect corresponding to the predetermined drop-down pattern of setting counting, as the drop-down pattern of the image of continuous input.
Program recorded medium is stored said procedure according to an embodiment of the invention.
In image processing apparatus according to an embodiment of the invention, among the image of input continuously, calculate before paying close attention to image and paying close attention to image and each of reference picture afterwards between correlation; Whether confirm to pay close attention to image corresponding to the interruption in the predetermined number of consecutive identical image based on correlation; If confirm to pay close attention to image corresponding to interruption, the continuous appearance of then counting identical image is as continuous occurrence count; If confirm to pay close attention to image, then confirm the setting counting that whether matees the indication continuous occurrence count corresponding in the continuous occurrence count of preceding image up to being right after with predetermined drop-down pattern corresponding to interruption; If confirm to pay close attention to image, and confirm then to count the number of times that occurs with predetermined drop-down pattern matched and count as coupling up to the continuous occurrence count match settings counting that is right after at preceding image corresponding to interruption; And if confirm to pay close attention to image corresponding to interruption; And confirm up to being right after at the continuous occurrence count match settings counting of preceding image and coupling counting then confirm corresponding to the predetermined drop-down pattern of setting counting it is the drop-down pattern of the image imported continuously greater than preset count.
Image processing apparatus can be an independent device according to an embodiment of the invention, perhaps can be the module of carrying out image processing.
According to embodiments of the invention, can come to detect drop-down pattern through single process configuration for the input picture of various frame frequencies, make it possible to thus to rebuild original progressive picture with high accuracy from input picture based on detected drop-down pattern.
Description of drawings
Fig. 1 shows the block diagram according to the configuration example of the image processing apparatus of the embodiment of the invention;
Fig. 2 is the state transition diagram that pictorial images is handled;
Fig. 3 is the flow chart of diagram progressive (progressive) image generative process;
Fig. 4 is the flow chart of diagram correlation value calculation process;
Fig. 5 is the figure of diagram correlation value calculation process;
Fig. 6 is the figure of diagram correlation value calculation process;
Fig. 7 is the flow chart of the drop-down pattern detection process of diagram;
Fig. 8 is the figure of the drop-down pattern detection process of diagram;
Fig. 9 is the figure that illustrates the setting of the predetermined threshold that is used for definite interruption (break); And
Figure 10 shows the figure of the configuration example of personal computer.
Embodiment
Before describing embodiments of the invention, the corresponding relation between the characteristic of the disclosed embodiments in the present invention and this specification is discussed below.Describe meaning and will guarantee in specification, to describe the support embodiments of the invention.Therefore, even the embodiment in the following description is not described to relate to a certain characteristic of the present invention, that neither mean that embodiment does not relate to this characteristic of the present invention.On the contrary, even embodiment is described to relate to a certain characteristic of the present invention, that neither mean that this embodiment does not relate to other characteristics of the present invention.
In addition, description is not the detailed description that meaning will provide all aspects of the present invention.That is, the still existence of the each side of the present invention of failed call protection in this application described in the specification is not denied in description, promptly possibly require protection through dividing an application or pass through the existence that modification requires the each side of the present invention of protection in addition in future.
Promptly; Image processing apparatus comprises according to an embodiment of the invention: the correlation value calculation parts (for example; Correlation value calculation part 13 among Fig. 1), be used for calculating at the concern image of the image of input continuously and before paying close attention to image and the correlation between each reference picture afterwards; Interrupt to confirm parts (for example, part 14 is confirmed in the interruption among Fig. 1), be used for confirming to pay close attention to image whether corresponding to the interruption of predetermined number of consecutive identical image based on correlation; Continuous occurrence count parts (the for example continuous occurrence count device 15 of the identical image among Fig. 1), if be used for confirming to pay close attention to image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count; Pattern confirms that parts (for example; Pattern among Fig. 1 is confirmed part 16); If be used for confirming to pay close attention to image, then confirm up to being right after in the continuous occurrence count of preceding image whether mate the setting counting of indication corresponding to the continuous occurrence count of predetermined drop-down pattern corresponding to interruption; The coupling counter block (for example; Match counter 19 among Fig. 1), if be used for confirming to pay close attention to image, and confirm up to the continuous occurrence count match settings counting that is right after at preceding image corresponding to interruption; Then count the number of times that occurs with predetermined drop-down pattern matched, count as coupling; And drop-down pattern detection parts (for example; Drop-down pattern detection part 18 among Fig. 1); Confirm to pay close attention to image corresponding to interruption if be used for; And confirm then to detect corresponding to the predetermined drop-down pattern of setting counting up to being right after at the continuous occurrence count match settings counting of preceding image and coupling counting greater than preset count, as the drop-down pattern of the image of continuous input.
Image processing apparatus can be configured to also comprise that progressive picture (for example generates parts; Progressive picture among Fig. 1 generates part 22); Be used for drop-down pattern, from the image generation progressive picture of continuous input based on the image of the continuous input that is detected by drop-down pattern detection parts.
Drop-down pattern detection parts (for example; Drop-down pattern detection part 18 among Fig. 1) can be configured feasible: if confirm to pay close attention to image corresponding to interruption; If confirm up to the continuous occurrence count match settings counting that is right after at preceding image; And the coupling counting is not more than preset count, but greater than another preset count littler than preset count, then drop-down pattern detection parts suspend the detection to predetermined drop-down pattern.
Interrupt confirming parts (for example, part 14 is confirmed in the interruption among Fig. 1) can be arranged to based on correlation whether whether confirm to pay close attention to image corresponding to the interruption in the predetermined number of consecutive identical image greater than the comparative result between predetermined threshold or the correlation.
Image processing apparatus can be configured to also comprise threshold value management component (for example, the threshold value administrative section 14a among Fig. 1), is used for managing predetermined threshold so that become littler according to continuous occurrence count.
The correlation value calculation parts (for example; Correlation value calculation part 13 among Fig. 1) each is paid close attention to image and is paying close attention to before the image and the correlation between the pixel near the first area the center of reference picture afterwards in the continuous image of importing can be arranged to calculating; And can detect under the state of predetermined drop-down pattern at drop-down pattern detection parts, be used for calculating the image of input continuously greater than the first area and before each is paid close attention to image and is paying close attention to image and the correlation between the pixel near the second area the center of reference picture afterwards.
Image processing method comprises step according to an embodiment of the invention: each that calculate in the continuous image of importing paid close attention to image and paid close attention to before the image and the correlation (for example, step S22 among Fig. 4 and S23) between the reference picture afterwards; Whether confirm to pay close attention to image corresponding to the interruption in the predetermined number of consecutive identical image (for example, step S42 among Fig. 7 and S44) based on correlation; If confirm to pay close attention to image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count (for example, the step S45 among Fig. 7); If confirm to pay close attention to image, then confirm the setting counting (for example, the step S48 among Fig. 7) that whether matees the indication continuous occurrence count corresponding in the continuous occurrence count of preceding image up to being right after with predetermined drop-down pattern corresponding to interruption; If confirm to pay close attention to image, and confirm then to count the number of times that occurs with predetermined drop-down pattern matched and count (for example, the step S51 among Fig. 7) as coupling up to the continuous occurrence count match settings counting that is right after at preceding image corresponding to interruption; And if confirm to pay close attention to image corresponding to interruption; And confirm to count greater than preset count up to the continuous occurrence count match settings counting and the coupling that are right after at preceding image; Then detect corresponding to the predetermined drop-down pattern of setting counting; Drop-down pattern (for example, the step S57 among Fig. 7) as the image of continuous input.
Fig. 1 shows image processing apparatus according to an embodiment of the invention.
Image processing apparatus 1 among Fig. 1 detects drop-down pattern from the mobile image that forms predetermined frame frequency, and further generates progressive picture based on detected drop-down pattern, and is being waited this progressive picture of demonstration on the display part 2 that forms by LCD (LCD).
Buffer 11-1 stores input picture to 11-3 by the place in order, and input images stored is supplied to the progressive picture generation part 22 of buffer 11-2 to 11-3 and downstream.Through this processing, in 11-3, accumulate three continous field images all the time at buffer 11-1.Should be noted that as field picture top field image and bottom field image be by alternately supply in order, and top field image and bottom field image form single two field picture.
Extract part 12 and be extracted in buffer 11-1 pixel value of the pixel near the zone the core of each field picture of storage respectively in the 11-3, and this pixel value is supplied to correlation value calculation part 13.In addition, extract part 12, near the scope in the extraction zone the part of main dispatching centre based on state information from 20 supplies of condition managing part.Should be noted that the detailed description that will provide state information after a while, and based near the regional scope of the extraction the core of state information.
Correlation value calculation part 13 is based near the pixel value of the pixel the center of three field picture of supplying from extraction part 12; Its that stored among the field picture that calculating is stored in buffer 11-2 (as paying close attention to field picture) and buffer 11-1 and the 11-3 before and each field picture afterwards (hereinafter; Be also referred to as the reference field image) between the summation of absolute difference of pixel at same position place; As correlation, and the correlation that calculates is supplied to interrupts confirming part 14.More specifically, because concern field picture and reference field image are respectively top field image and bottom field image, so their location of pixels differs from one another.Thereby; Based on each pixel of paying close attention to the field; The pixel at 13 pairs of same position places of correlation value calculation part is carried out interpolation, each pixel of field picture as a reference, and the pixel of utilizing interpolation and each respective pixel, through use absolute pixel differences and calculate correlation.In addition, correlation value calculation part 13 also the pixel through using the same position place between reference field absolute pixel differences with, calculate different correlations.
Interrupt to confirm part 14 relatively from the correlation and the threshold value of 13 supplies of correlation value calculation part, confirm that thus whether concern field picture as processing target is corresponding to interruption.If detect interruption, interrupt then confirming that part 14 has detected the information of interruption to the 16 notice indications of pattern detection part.If do not detect interruption; Then interruption confirms that part 14 makes the continuous occurrence count device 15 of identical image upwards count indication and detects and the continuous occurrence count that is right after in the continuous occurrence number of the identical image of preceding field picture, and when detecting interruption, continuous occurrence count is resetted.In addition, interrupt confirming that part 14 comprises threshold value administrative section 14a.Threshold value administrative section 14a comes change threshold according to the continuous occurrence count of storage in the continuous occurrence count device 15 of identical image.
At this, interruption will be described.For example, drop-down if drop-down pattern is 2-3, the homologous field image is by twice, three times, twice of repetition alternately, three inferior.At this moment; When correlation hour; Interrupt definite part 14 this is used as indication and the continuous appearance that is right after in the consistent field picture of preceding field picture; And when correlation is big, interrupt confirming that field picture that part 14 locates tip position (at the tip position place, two or three homologous field image switchings of group are to next two or three field picture organizing continuously continuously) is as interruption.Thereby; At drop-down pattern is under the drop-down situation of 2-3; Occur continuously detecting after twice corresponding to the field picture of interrupting at the homologous field image; And occurring continuously detecting once more corresponding to the field picture of interrupting after three times, and repeat identical processing thereafter from above-mentioned field picture, homologous field image.
Notified through after interrupting confirming that part 14 detects the fact corresponding to the field picture of interrupting; Pattern confirms that part 16 is based on up to the continuous occurrence count that is right after in preceding field picture; Whether corresponding continuous occurrence count of continuous occurrence count and the above-mentioned continuous occurrence count of confirming the drop-down pattern of storage in indication and the drop-down styles store part 17 mate, and confirm thus to be right after whether mating drop-down pattern corresponding to the continuous occurrence number of the homologous field image before the field picture of interrupting.
If pattern is confirmed that part 16 is confirmed to be right after and is being mated drop-down pattern corresponding to the continuous occurrence number of the homologous field image before the field picture of interrupting; Then pattern confirms that part 16 makes match counter 19 storages about mating the information of counting, the number of this coupling counting indication and drop-down pattern matched.In addition, pattern confirms that part 16 notifies drop-down pattern detection part 18 to detect and the continuous fact of the drop-down pattern of occurrence count coupling.That is, for example, in advance various drop-down patterns are registered in the drop-down styles store part 17, certainly, it is drop-down to register 2-3 in advance.Thereby when detecting continuous occurrence count when interrupting and be 2 or 3, pattern confirms that part 16 is used as this as the indication match patterns and exists, and make match counter 19 upwards counting about the drop-down coupling counting of 2-3.Yet, should be noted that in continuous occurrence count be under 2 the situation, if for example the drop-down grade of 2-2 is registered as drop-down pattern in advance, possibly be difficult to identify continuous which drop-down pattern of occurrence count indication.Therefore, upwards counting should the coupling counting to two drop-down patterns through match counter 19, and will mate counting with just before the drop-down pattern of registering compare, definite thus drop-down pattern is drop-down pattern of 2-2 or the drop-down pattern of 2-3.In drop-down styles store part 17, except 2-3 is drop-down, store the various drop-down pattern drop-down such as 8-7, that 6-4 is drop-down and 5-5 is drop-down in advance.
Notified when detecting drop-down pattern true through pattern detection part 16; Drop-down pattern detection part 18 inquiry match counters 19 are to read the information about the number of times that has mated to each drop-down pattern; I.e. coupling counting, and about the information of corresponding drop-down pattern.When the number of times that has mated during greater than pre-determined number, drop-down pattern detection part 18 is used as corresponding drop-down pattern as the drop-down pattern of input picture, and the drop-down pattern of the input picture that is detected is supplied to progressive picture generates part 22.
Condition managing part 20 is according to extracting part 12, correlation value calculation part 13, interrupting confirming that part 14, pattern confirm each mode of operation of part 16 and drop-down pattern detection part 18; Confirm that through making mode of operation be converted to the state that do not detect shown in Figure 2, pattern detection state, pattern state, drop-down pattern status recognition and halted state come the bookkeeping state, confirm that thus drop-down pattern is drop-down pattern of 2-2 or the drop-down pattern of 2-3.In drop-down styles store part 17, except 2-3 is drop-down, store the various drop-down pattern drop-down such as 8-7, that 6-4 is drop-down and 5-5 is drop-down in advance.
The state of not detecting refers to the state of at all finding interruption.When detecting interruption, do not detect state-transition to the pattern detection state.
The pattern detection state refers to has no progeny, does not detect the state that next interrupts and therefore also do not detect the pattern of forming drop-down pattern in detecting for the first time.In detecting, have no progeny, the pattern detection state-transition is confirmed state to pattern.
Pattern confirms that state refers to the continuous occurrence count that is used to be based on two homologous field images between the interruption and matees the state that the number of times of drop-down pattern (i.e. coupling counting) confirms to form the pattern of drop-down pattern.When coupling was counted greater than higher limit, state-transition was to drop-down pattern status recognition.When coupling was counted less than lower limit, state-transition was to the pattern detection state.When coupling was counted less than higher limit and greater than lower limit, state was converted to the pattern detection state via halted state.When continuous occurrence count did not match any drop-down pattern, state-transition was to not detecting state.
Based on the drop-down pattern from drop-down pattern detection part 18 supplies, progressive picture generates the input picture that part 22 conversions are supplied from buffer 11-3 in order, is used for the progressive picture that on display part 2, shows with generation.
Next, with reference to the flow chart of figure 3, with the description that provides the image generative process that realizes by the image processing apparatus among Fig. 1.
At step S1, buffer 11-1 determines whether to have supplied field picture to 11-3, and repeats same deterministic process up to having supplied field picture.In step S1, for example,, then handle and advance to step S2 if supplied field picture in order.
At step S2, buffer 11-1 stores the field picture of supply in order to 11-3, and previously stored field picture is supplied to the downstream.That is, field picture that buffer 11-1 storage has been imported, and put up with the field picture of having stored before this simultaneously and be supplied to buffer 11-2.Buffer 11-2 storage is from the field picture of buffer 11-1 supply, puts up with the field picture of having stored before this simultaneously and is supplied to buffer 11-3.Buffer 11-3 storage is from the field picture of buffer 11-2 supply, puts up with the field picture of having stored before this simultaneously and is supplied to progressive picture and generates part 22.Should be noted that owing to when handling for the first time, do not have field picture to be provided to buffer 11-2 and 11-3, therefore after having supplied at least three field picture, begin main processing.
At step S3, extract part 12 and carry out the correlation value calculation processes, to calculate correlation, as the correlation of paying close attention to field picture as field picture processing target, that store among the buffer 11-2 with correlation value calculation part 13.
Now, with reference to the flow chart of figure 4, with the description that provides the correlation value calculation process.
At step S21, extract part 12 inquiry condition managing parts 20 to confirm whether current state is drop-down pattern status recognition.For example; If condition managing part 20 inquiry state storage parts 21 are with read status information; And the state information that so reads indication current state is drop-down pattern status recognition; Then in step S22, extract part 12 and be extracted in from buffer 11-1 to 11-3 and pay close attention near the regional interior pixel bigger field picture and the center of paying close attention to before the field picture with field picture afterwards, and the pixel of being extracted is supplied to correlation value calculation part 13 than predetermined value.
On the other hand; For example; If state information indication current state is not drop-down pattern status recognition; Then in step S23, extract part and be extracted in from buffer 11-1 to 11-3 and pay close attention near the regional interior pixel littler field picture and the center of paying close attention to before the field picture with field picture afterwards, and the pixel of being extracted is supplied to correlation value calculation part 13 than predetermined value.
That is,, this means also successfully not detect drop-down pattern if current state is not drop-down pattern status recognition.Thereby, for example as shown in Figure 5, be extracted near the center of field picture P less than the pixel in the zone of the a1 * a2 of predetermined value.This is owing to following reason.That is, in the zone at the edge that more approaches field picture P, often comprise that like captions (telops), this possibly cause being difficult to confirming identical image by an image that changes.Thereby, near the center of field picture P, be read to obtain correlation less than the pixel in the zone of predetermined value.On the other hand, if current state is drop-down pattern status recognition, this means successfully to detect drop-down pattern.Thereby, for example as shown in Figure 5, be extracted near the center of field picture P b1 * b2 greater than predetermined value (b1＞a1, the pixel in the zone of b2＞a2).This is because owing to might detect drop-down pattern, can not cause being difficult to confirming identical image so expection enlarges the extraction zone.Thereby the center that reads in field picture P is neighbouring greater than the pixel in the zone of predetermined value.
At step S24, correlation value calculation part 13 obtain absolute pixel differences between the pixel of being supplied with, obtaining the correlation of paying close attention to field picture, and this correlation be supplied to interrupt confirming part 14.More specifically; Before concern field picture of in buffer 11-2, storing and the concern field picture of in buffer 11-1 and 11-3, storing with field picture afterwards (reference field image) between relation make: if the concern field picture is a top field; Be bottom field with afterwards field picture then; On the contrary, be bottom field if pay close attention to field picture, then before with afterwards field picture be top field.Therefore, correlation value calculation part 13 is through using before and each pixel of reference field image afterwards, to carrying out virtual interpolation with the corresponding pixel of location of pixels of paying close attention to field picture.
Promptly; For example; As shown in Figure 6; Field picture when supposing to pay close attention to field picture and be tb constantly, the pixel on the vertical direction are pixel P11 and the P12 that the bottom from figure begins to count, before and reference field image afterwards be respectively the field picture when moment ta, tc; And their each pixel is pixel P1 and pixel P21, then on correlation value calculation part 13 each field picture when moment ta and tc to carrying out interpolation with the corresponding pixel P11 ' of location of pixels that pays close attention to field picture (=(P12+P11)/2).In addition; Correlation value calculation part 13 obtain the pixel of being obtained 11 ' and the absolute value of the difference between each of the pixel P11 of the corresponding position on the reference field image and P12 (| P11 '-P1| and | P11 '-P21|); And that obtains in the same manner is calculated as correlation PR and AR corresponding to the absolute difference of each pixel with (that is, absolute difference with).In addition, obtain absolute difference between pixel P1 and P21 (| P1-P21|), and obtain in the same manner be calculated as correlation FR corresponding to the absolute difference of each pixel with (that is, absolute difference with).
Should be noted that in Fig. 6 trunnion axis representes to indicate the demonstration time orientation constantly of each field picture, vertical axis is represented the location of pixels of each field picture on vertical direction (x direction).
The correlation FR that so obtains make when be right after preceding image or be right after after the identical image of image when occurring continuously; The absolute difference of the pixel value between the pixel with become less; And when identical image does not occur continuously, the absolute difference of the pixel value between the pixel with become bigger.That is, if pay close attention to field picture with preceding and after the reference field picture altitude relevant, then correlation FR becomes little value, and correlation PR and AR become and be roughly the same.On the contrary, if pay close attention to field picture with preceding and after the reference field image correlation lower, for example, if pay close attention to field picture with lower at preceding reference field image correlation, then correlation FR becomes big value, and correlation AR becomes less with respect to correlation PR.
Calculate correlation through said process, and it is supplied to the definite part 14 of interruption.
Now, description will be got back to the flow chart of Fig. 3.
When in step S3, calculating correlation, in step S4, carry out drop-down pattern detection process, and detect drop-down pattern.
Now, with reference to the flow chart of figure 7, drop-down pattern detection process will be described.
At step S41, interrupt confirming part 14 to condition managing part 20 requesting state information, and, confirm that current state is pattern detection state or drop-down pattern detection state based on the state information of being obtained.For example,, but do not detect state, then handle advancing to step S42 if be not pattern detection state or drop-down pattern detection state by condition managing part 20 states management and that be stored in the state storage part 21.
At step S42, interrupt confirming correlation PR, AR and the FR of part 14 based on 13 supplies of autocorrelation value calculating section, confirm whether to pay close attention to field picture corresponding to interruption.For example, at step S42, if confirm that correlation FR is less than predetermined threshold " th "; Perhaps correlation FR is greater than predetermined threshold " th "; Correlation AR is not less than correlation PR, that is, it is consistent with the field picture that is right after preceding to pay close attention to field picture; Then processing finishes, state without as by the indicated transformation of transformation C1 among Fig. 2.
That is, for example as shown in Figure 8, start from processing for preceding several field picture, can not in drop-down pattern, identify the position of these field picture, so state remaining the state of not detecting under the situation of moment t0.
Should be noted that in Fig. 8, trunnion axis express time axle, and vertical axis is illustrated in that each is located constantly, pixel vertical arrangement on the given horizontal coordinate in field picture.That is, in Fig. 8, handle beginning at moment t0 place, and by the circle indication moment t1 during to t16 in field picture accordingly the pixel of vertical arrangement.Constantly show when t1, t3, t5, t7, t9, t11, t13 and t15 at example at each, and show when t2, t4, t6, t8, t10, t12, t14 and t16 constantly at example corresponding to the vertical arrangement of two pixels in the field picture of bottom field (BottomField) at each corresponding to the vertical arrangement of two pixels in the field picture of top field (TopField).In addition, the pixel value of each pixel is indicated by the pattern in the circle of indication pixel.Respectively moment t1 in the regional Z1 of moment t3, constantly t4 in the regional Z2 of t8, constantly t9 in the regional Z3 of t13 and constantly t13 the circle, white circle of band point, the tiltedly circle of shade and the circle of grid have been shown in the regional Z4 of t16.The identical identical pixel value of pattern indication.
Therefore; For example; If the field picture when paying close attention to field picture and be t2 constantly is as shown in Figure 8, then constantly the pixel value of the pixel of the field picture during t2 with when the moment t1 be right after at the pixel value of the pixel of preceding field picture and during at moment t3 be right after after the pixel value of pixel of field picture identical.Therefore, correlation FR is still less, and does not become greater than predetermined threshold " th ", so state remains the state of not detecting.
If the field picture when for example paying close attention to field picture and be t3 constantly, as shown in Figure 8, then the pixel value of the pixel of the field picture when moment t4 is different from the pixel value of the pixel of the field picture when moment t2, so correlation FR becomes greater than predetermined threshold " th ".Yet; The pixel value of the pixel of the field picture when moment t3 being right after at the pixel value of the pixel of preceding field picture during with t2 constantly is identical; And during with moment t4 be right after after the pixel value of pixel of field picture different; Therefore correlation PR trends towards diminishing, and correlation AR trends towards becoming greatly.Therefore, correlation AR is not less than correlation PR, so state remains the state of not detecting.
On the other hand, in step S42, if for example correlation FR is greater than predetermined threshold " th "; And correlation AR is less than correlation PR; That is, be different from the field picture that is right after preceding, and therefore think and detect interruption if pay close attention to field picture; Then in step S43, interrupt confirming that part 14 notify status administrative sections 20 have detected the fact of interruption.Handle in response to this, shown in the transformation C2 among Fig. 2, it is the pattern detection state that condition managing part 20 state that makes never detects state-transition, and with state storage in state storage part 21.
Promptly; The field picture when if concern field picture for example as shown in Figure 8 is moment t4; For example, when moment t5 be right after after the pixel of field picture do not have being right after with the time at moment t3 at the identical pixel value of the pixel of preceding field picture, so correlation FR becomes greater than predetermined threshold " th ".In addition, the being right after when pixel of the field picture during owing to moment t4 does not have with moment t3 at the identical pixel value of the pixel of preceding field picture, so correlation PR trends towards becoming big.And, since the pixel of field picture when moment t4 when having with moment t5 be right after after the identical pixel value of pixel of field picture, so correlation AR trends towards diminishing.Therefore, correlation AR becomes less than correlation PR, thereby can identify and pay close attention to the tip position that field picture is positioned at one group of field picture of importing continuously, and therefore corresponding to interruption.
Through this processing, as shown in Figure 2, state-transition is to the pattern detection state of the tip position of the field picture that identifies continuous input, therefore when later homologous field image without interruption, can begin the counting of continuous occurrence count C.
On the other hand; If confirm that at step S41 state is for example pattern detection state or drop-down pattern detection state; Then in step S44; As in the processing of step S42, interrupt confirming correlation PR, AR and the FR of part 14 based on 13 supplies of autocorrelation value calculating section, confirm whether to pay close attention to field picture corresponding to interruption.
Do not correspond to interruption if confirm to pay close attention to field picture at step S44; Promptly; If imported continuously and be right after, then in step S45, interrupt confirming that part 14 increases progressively 1 with the continuous occurrence count C of storage in the continuous occurrence count device 15 of identical image in the identical field picture of preceding field picture.At this moment, as by shown in the transformation C3 or C10 among Fig. 2, state-transition not taking place, only continuous occurrence count is increased progressively 1.Therefore, to t8 or when constantly t9 is to t13, continuous occurrence count is increased progressively 1 at moment t5 shown in Figure 8.
In step S46, the threshold value administrative section 14a that interrupts confirming part 14 is provided with predetermined threshold " th " by mode for example shown in Figure 9 according to continuous occurrence count C, and processing finishes.That is, under for example not having continuously the situation that very state of about-face recurs in the field picture of input, if in the case predetermined threshold " th " is set to constantly, the state that then can not detect interruption possibly continue.Therefore, as shown in Figure 9, therefore when correlation " dif " has become less than predetermined value " dif1 " and think that image does not have that very the state of about-face has continued, reset predetermined threshold " th " according to correlation " dif ".Even this processing makes when the image that has only very little change occurs continuously, also detects interruption more easily.
Should be noted that in Fig. 9 trunnion axis is represented correlation " dif ", and vertical axis is represented predetermined threshold " th ".
If the difference between the field picture is somewhat little, then expection is paid close attention to field picture and is not corresponded to interruption.Yet, depend on employed design pattern, always do not calculate big correlation.Thereby, be greatly or little in order to confirm correlation, can use before to be used for confirming the correlation that interrupts.That is, suppose that generally correlation will reduce when scene is static.Thereby, can the correlation when confirming come relatively to confirm that correlation is greatly or little, thereby confirm interruption with the interruption that obtained last time.
On the other hand, if in step S44, confirm to pay close attention to field picture corresponding to interruption, then in step S47, condition managing part 20 makes state-transition to the pattern detection state, and with state storage in state storage part 21.Handle through this, as indicated by transformation C4 or C9 among Fig. 2, state-transition is confirmed state to pattern.
At step S48, pattern confirm continuous occurrence count C that part 16 confirms the continuous occurrence count devices 15 of identical image whether with set count matches.The continuous occurrence number that refers to the homologous field image of forming drop-down pattern " sets and count " to term in this use.For example, Fig. 8 shows the example of the field picture of under the drop-down situation of 5-5, importing, and has wherein imported five homologous field images continuously from moment t4 to t8 with from moment t9 to t13, and the expression one-time continuous has been imported five identical field picture.Therefore, if imported five homologous field images continuously, then this can be confirmed as the quantity corresponding to the continous field image of forming the drop-down pattern of 5-5.Because continuously occurrence count C representes to be confirmed as the occurrence number of the field picture identical with the field picture at the top of continous field image, if therefore imported five homologous field images, the continuous occurrence count C=4 of continuous occurrence count device 15 indications of identical image then.Therefore, under the situation of the drop-down pattern of 5-5, setting counting is 4.
On the contrary, under the drop-down situation of 2-3, the homologous field image appears in the sequence of 2 field picture, 3 field picture etc., and setting counting is 1 and 2.In addition, under the drop-down situation of 8-7, the homologous field image appears in the sequence of 8 field picture, 7 field picture etc., and therefore setting counting is 7 and 6.Under the drop-down situation of 6-4, the homologous field image appears in the sequence of 6 field picture, 4 field picture etc., and therefore setting counting is 5 and 3.
Thereby at step S48, under the situation of Fig. 8, continuously occurrence count C becomes 4 when moment t9, therefore confirm continuous occurrence count with set count matches, and handle and advance to step S49.
At step S49, pattern confirms that part 16 estimations have the drop-down pattern of the setting counting of coupling.That is, pattern confirms that part 16 estimates corresponding drop-down pattern from counting with the setting of continuous occurrence count C coupling.Under the situation of Fig. 8, occurrence count is 4 continuously, estimates that therefore the drop-down pattern of coupling is that 5-5 is drop-down.
At step S50, pattern confirm part 16 read with match counter 19 in the drop-down pattern stored explicitly of the coupling counting M of storage, and confirm whether the drop-down pattern that reads matees estimated drop-down pattern.That is, because the drop-down pattern quilt of coupling and coupling counting M be stored in the match counter 19 explicitly, so pattern confirms drop-down pattern and estimated drop-down pattern that part 16 is relatively stored, with definite both whether mate.
If drop-down pattern of in step S50, confirming to store explicitly and estimated drop-down pattern matched with coupling counting M; Then in step S51; Pattern confirms that part 16 makes match counter 19 will indicate the coupling counting M with the quantity of setting count matches to increase progressively 1; And store resulting coupling counting M explicitly, and notify drop-down pattern detection part 18 to take place and set the fact of count matches with estimated drop-down pattern.
Thereby, under the situation of Fig. 8, when moment t9, increase progressively 1 for the drop-down counting M that will mate of 5-5.And if for example continuous occurrence count C is 7 or 6, then pattern confirms that part 16 increases progressively 1 for the drop-down counting M that will mate of 8-7.In addition, if for example continuous occurrence count C is 5 or 3, then pattern confirms that part 16 increases progressively 1 for the drop-down counting M that will mate of 6-4.
On the other hand; If drop-down pattern of confirming to store explicitly with coupling counting M at step S50 and estimated drop-down pattern do not match; Then pattern confirms that part 16 will mate counting M and reset to 1, and registers resulting coupling counting M explicitly with estimated drop-down pattern.That is, changed before the detection of interrupting with after detecting with coupling counting M drop-down pattern of storing explicitly and the drop-down pattern that the estimated unmatched fact of drop-down pattern means very possible input picture, therefore coupling is counted and is reset.
At step S53, drop-down pattern detection part 18 confirms that whether coupling counting M is less than lower limit.If coupling is counted M less than lower limit, then at step S54, drop-down pattern detection part 18 notify status administrative sections 20 these results.Therefore, condition managing part 20 makes state-transition to the pattern detection state.That is, shown in the transformation C5 among Fig. 2, state confirms that from pattern state-transition is to the pattern detection state.
On the other hand, if coupling counting M is not less than lower limit in step S53, then in step S55, drop-down pattern confirms that whether part 18 definite coupling counting M are greater than higher limit.If coupling counting M is greater than higher limit; Then drop-down pattern confirms that part 18 thinks that the continuous occurrence count C of this indication homologous field image has repeated to occur repeatedly with the coupling of setting counting; Make it possible to detect drop-down pattern, and notify status administrative section 20 these results.
In step S56, condition managing part 20 state that makes confirms that from pattern part is converted to drop-down pattern status recognition.Handle through this, indicated like the transformation C8 among Fig. 2, state confirms that from pattern state-transition is to drop-down pattern status recognition.
In step S57; Drop-down pattern detection part 18 will about with match counter 19 in the information of the drop-down pattern stored explicitly of the coupling counting M of storage be supplied to progressive picture and generate part 22, as information about the drop-down pattern that detects from input picture.
In step S58, drop-down pattern detection part 18 resets the continuous occurrence count C of storage in the continuous occurrence count device 15 of identical image, and processing finishes.
On the other hand; If the coupling counting is not more than higher limit in step S55; That is, if coupling counting M greater than lower limit less than higher limit, and the number of times of the homologous field image of the drop-down pattern of composition that has been consecutively detected is less; Then in step S59, drop-down pattern detection part 18 notify status administrative sections 20 coupling counting M greater than lower limit less than the fact of higher limit.In response to this notice, condition managing part 20 make state from the pattern detection state-transition to halted state.That is, through the transformation C6 among Fig. 2, state from the pattern detection state-transition to halted state.
In step S60, drop-down pattern detection part 18 resets to 0 with the continuous occurrence count C of the continuous occurrence count device 15 of identical image, and this result of notify status administrative section.
In step S61, based on this notice, condition managing part 20 makes state-transition to the pattern detection state.That is, as by shown in the transformation C7 among Fig. 2, state is converted to the pattern detection state from halted state.
In addition; In step S48; If the continuous occurrence count C of the continuous occurrence count device 15 of identical image does not match with the setting counting; That is,, then can not detect said drop-down pattern if continuous occurrence count C does not correspond to the setting counting of the homologous field image of forming the drop-down pattern of registering in advance in the drop-down styles store part 17.Thereby; In step S62; Pattern is confirmed part 16 with the continuous occurrence count C of storage in the continuous occurrence count device 15 of identical image, the coupling counting M of match counter and all initialization of predetermined threshold " th " of being managed by the threshold value administrative section 14a that interrupts in definite part 14, and notify status administrative section 20 these results.
In step S63, the state information of storage in the condition managing part 20 reading state storage areas 21, and the state that makes confirms that from pattern state-transition is to not detecting state.That is, through the transformation C11 among Fig. 2, state confirms that from pattern state-transition is to not detecting state.
Now, description will be got back to the flow chart of Fig. 3.
When in step S4, accomplishing drop-down pattern detection process, at step S5, progressive picture generates part 22 and determines whether to detect drop-down pattern.That is,, therefore always do not detect drop-down pattern, should handle only if in the sequence of step S41, S44, S47 and S58, carry out because drop-down pattern detection process is the process that each field picture is carried out.
Then, if in step S5, detect drop-down pattern, then in step S6, based on drop-down pattern, progressive picture generates part 22 will convert the progressive picture that is used for 2 demonstrations in the display part to from the input picture of buffer 11-3 supply.
On the other hand, if do not detect drop-down pattern, then skips steps S6, and processing gets back to step S1, and repeats processing subsequently.
Above-mentioned processing can be summarized as follows.That is, for example, under the situation of Fig. 8, when t0 began to detect in the moment, before moment t3, state remained the state of not detecting, and repeated among Fig. 7 step S41 to the processing of S43.When detecting interruption at moment t4, state-transition is to the pattern detection state.During the period from moment t4 to t8, the continuous occurrence count C of storage increases progressively in order in the continuous occurrence count device 15 of the processing of repeating step S41, S44 and S45, so identical image in order.When the moment that detects interruption at moment t9 once more, state-transition is confirmed state to pattern.At this moment and since continuously occurrence count C with set counting 4 couplings, handle as first, will mate and count M=1 and 5-5 is drop-down is stored in explicitly in the match counter 19.
In addition, suppose that lower limit is 1, and higher limit is 2, then after state temporarily is set to halted state, state-transition is to the pattern detection state, and during the period from moment t10 to t13, the processing of repeating step S41, S44 and S45.Then, when detecting interruption once more at moment t4, state-transition is confirmed state to pattern.At this moment, continuously occurrence count once more with set counting 4 couplings, therefore increase progressively coupling counting M, and will mate and count M=2 and 5-5 is drop-down is stored in explicitly in the match counter 19.
When coupling counting M becomes 2 and during therefore greater than higher limit, the drop-down pattern that drop-down pattern detection part 18 will be stored with the coupling counting M of storage in the match counter 19 explicitly is supplied to progressive picture and generates part 22, as the drop-down pattern of input picture.
Behind moment t15, as long as identical drop-down pattern continues, when detecting interruption, coupling counting M just continues to become bigger, therefore continues the drop-down pattern of detection at every turn.
In addition, even when drop-down pattern changes midway, the time point place that surpasses higher limit at coupling counting M detects new drop-down pattern once more.That is, even under the situation of the image that various drop-down patterns dynamically change, also can continue to detect drop-down pattern.
Although aforementioned description be directed to the absolute pixel differences used between the image and as correlation, correlation can be the arbitrary value of the relevance between can presentation video, and therefore also can use other correlations.For example, can use the stream header information of the encoded image sequence before decoded to be used as correlation.
According to above-mentioned configuration; Even under the situation of the interlaced image of creating from the progressive picture of low frame rate (interlacedimage); It or not the image of inserting in creating through interpolation; And can come to rebuild fully original progressive picture based on the drop-down pattern that detects, make it possible to realize the improvement of picture quality thus.
Even because when frame frequency changes, also can continue the detection of drop-down pattern, so can be suppressed at the deterioration of image that takes place during the period till detecting drop-down pattern once more after the detection of the drop-down pattern of temporary transient completion.
In addition, change or possibly carry out editor's occasion,, then suspend and handle if drop-down modal shift and proper process become very difficult in scene.Therefore, can suppress such as deterioration of image quality as the broach phenomenon that when different field picture rebuilds original progressive picture by error, takes place (combing phenomenon).
In addition; Even there is not very about-face in the original image; To such an extent as to and very little feasible be difficult to and noise is distinguished of the correlation in the design pattern, also be detected as the number of identical continuous appearance and dynamically change about the predetermined threshold of correlation according to field picture wherein.Therefore, even for the small correlation of representing animation (animation), also can continue to detect with good accuracy.In addition, can suppress owing to general under the situation of minor variations observed broach or Interpolation Process cause deterioration of image quality, realize the improvement of picture quality thus.
In addition, even, therefore when design pattern changes after this, also can keep detection owing to when in original image, good correlation occurring, also can not keep detection.
Because minimum three field picture are enough carried out all drop-down detections, therefore can minimize the capacity of frame memory.Be directed to the situation of using three field picture although should be noted that aforementioned description, also can use the field picture of greater number.In addition, can use two field picture to replace field picture.
In addition, because a plurality of drop-down patterns of one-time detection, therefore need be for each drop-down pattern does not provide treatment system or circuit, thereby allow to reduce throughput or circuit scale.
According to above-mentioned configuration, can come detect drop-down pattern through the configuration of single process for the input picture of various frame frequencies, make it possible to rebuild original progressive picture thus with high accuracy based on detected drop-down pattern and from input picture.
Although above-mentioned a series of process can be carried out through hardware, also can pass through software executing.If will be by a series of processes of software executing, the program that then will form software be installed to the computer that embeds the proprietary hardware from program recorded medium, for example when being mounted program, can carry out the general purpose personal computer of various functions etc.
Figure 10 shows the configuration example of all-purpose computer.Personal computer has built-in CPU (CPU) 1001 wherein.Input/output interface 1005 is connected to CPU 1001 via bus 1004.ROM (read-only memory) 1002 and RAM (random access memory) 1003 are connected to bus 1004.
Be connected to input/output interface 1005 be the importation 1006 that forms by the input equipment that uses its input operation order such as users such as keyboard or mouses, to display device output handle the output 1007 of function screen and result image, the storage area 1008 that forms by the hard drive of stored program and various times etc. and form and via the communications portion 1009 of the network executive communication process of representing by the internet by LAN (local area network (LAN)) adapter etc.Be also connected to input/output interface 1005 be from/to driver 1010 such as the removable media 1011 read/write data of disk (comprising floppy disk), CD (comprising CD-ROM (compact-disc-read-only memory) and DVD (digital universal disc)), magneto optical disk (comprising MD (mini-disk)) or semiconductor memory.
It should be noted that; In this specification; The step that description is recorded in the program on the recording medium not only comprises in time the order that in description, occurs with them and the process of carrying out in order, and comprises and be not to carry out in order on the time but process parallel or that carry out independently.
It will be appreciated by those skilled in the art that as long as in the scope of accompanying claims or its equivalent, depend on design demand and other factors, various modifications, combination, son combination and change can take place.
The cross reference of related application
The present invention comprises the Japanese patent application JP 2007-173590 theme of submitting in Japan Patent office with on August 10th, 2007 and on January 31st, 2008 relevant with JP 2008-020335, by reference its full content is herein incorporated.
Claims (7)
1. image processing apparatus comprises:
The correlation value calculation parts are used for calculating image in input continuously, are paying close attention to image and correlation between said concern image each before and in the reference picture afterwards;
Interrupt to confirm parts, be used for confirming that based on said correlation said concern image is whether corresponding to the interruption of predetermined number of consecutive identical image;
Continuous occurrence count parts, if be used for confirming said concern image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count;
Pattern is confirmed parts, if be used for confirming said concern image corresponding to interruption, then confirms up to being right after in the continuous occurrence count of preceding image whether mate indication and the setting counting of being scheduled to the corresponding continuous occurrence count of drop-down pattern;
The coupling counter block; If be used for confirming that said concern image is corresponding to interruption; And confirm to mate said setting counting, then to counting, as the coupling counting with the number of times of said predetermined drop-down pattern matched appearance up to being right after in the continuous occurrence count of preceding image; And
Drop-down pattern detection parts; If be used for confirming that said concern image is corresponding to interruption; And confirm to mate said setting counting and said coupling counting greater than preset count up to being right after in the continuous occurrence count of preceding image; Then detect predetermined drop-down pattern, as the drop-down pattern of the image of said continuous input corresponding to said setting counting.
2. according to the image processing apparatus of claim 1, also comprise:
Progressive picture generates parts, is used for the drop-down pattern based on the image of the continuous input that is detected by said drop-down pattern detection parts, from the image generation progressive picture of said continuous input.
3. according to the image processing apparatus of claim 1, wherein:
If confirm that said concern image is corresponding to interruption; Mate said setting counting if then confirm up to being right after in the continuous occurrence count of preceding image; And said coupling counting is not more than preset count; But greater than another preset count littler than said preset count, then said drop-down pattern detection parts suspend the detection to said predetermined drop-down pattern.
4. according to the image processing apparatus of claim 1, wherein:
The definite parts of interruption whether greater than the comparative result between predetermined threshold or the correlation, confirm that whether said concern image is corresponding to the interruption in the predetermined number of consecutive identical image based on said correlation.
5. according to the image processing apparatus of claim 4, also comprise:
Threshold value management component is used for managing said predetermined threshold so that said predetermined threshold becomes littler according to said continuous occurrence count.
6. according to the image processing apparatus of claim 1, wherein:
Said correlation value calculation parts calculate in the image of said continuous input, the correlation between the pixel near the first area said concern image and each the center before said concern image and in the reference picture afterwards;
Can detect at said drop-down pattern detection parts under the state of said predetermined drop-down pattern; Said correlation value calculation parts calculate in the image of said continuous input, the correlation between the pixel near the second area said concern image and each the center before said concern image and in the reference picture afterwards, and wherein said second area is bigger than said first area.
7. image processing method comprises step:
Calculating in the image of continuously input, paying close attention to image and correlation between said concern image each before and in the reference picture afterwards;
Based on said correlation, confirm that whether said concern image is corresponding to the interruption in the predetermined number of consecutive identical image;
If confirm said concern image corresponding to interruption, the continuous appearance of then stored counts identical image is as continuous occurrence count;
If confirm said concern image, then confirm up to being right after in the continuous occurrence count of preceding image whether mate indication and the setting counting of being scheduled to the corresponding continuous occurrence count of drop-down pattern corresponding to interruption;
If confirm said concern image, and confirm to mate said setting counting, then to counting, as the coupling counting with the number of times of predetermined drop-down pattern matched appearance up to being right after in the continuous occurrence count of preceding image corresponding to interruption; And
If confirm that said concern image is corresponding to interruption; And confirm to mate said setting counting and said coupling counting greater than preset count up to being right after in the continuous occurrence count of preceding image; Then detect predetermined drop-down pattern, as the drop-down pattern of the image of said continuous input corresponding to said setting counting.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP208947/07 | 2007-08-10 | ||
| JP2007208947 | 2007-08-10 | ||
| JP020335/08 | 2008-01-31 | ||
| JP2008020335A JP5040687B2 (en) | 2007-08-10 | 2008-01-31 | Image processing apparatus and method, and program |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101365093A CN101365093A (en) | 2009-02-11 |
| CN101365093B true CN101365093B (en) | 2012-02-08 |
Family
ID=40391192
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008102102790A Expired - Fee Related CN101365093B (en) | 2007-08-10 | 2008-08-11 | Image processing apparatus and method |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP5040687B2 (en) |
| KR (1) | KR20090016383A (en) |
| CN (1) | CN101365093B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6557524B2 (en) * | 2015-06-22 | 2019-08-07 | 株式会社メガチップス | Receiving apparatus and image transmission system using the same |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001094950A (en) * | 1999-09-20 | 2001-04-06 | Matsushita Electric Ind Co Ltd | Video id signal processor |
| JP2005117202A (en) * | 2003-10-06 | 2005-04-28 | Oki Electric Ind Co Ltd | Method for converting interlace video image into progressive video image |
| JP4225930B2 (en) * | 2004-02-04 | 2009-02-18 | 三菱電機株式会社 | Video signal processing apparatus and video signal processing method |
| JP4293963B2 (en) * | 2004-09-27 | 2009-07-08 | 三菱電機株式会社 | Video signal processing device, video signal processing method, and video signal display device |
| JP4749314B2 (en) * | 2006-11-28 | 2011-08-17 | 富士通株式会社 | Interlace / progressive conversion method and conversion apparatus |
-
2008
- 2008-01-31 JP JP2008020335A patent/JP5040687B2/en not_active Expired - Fee Related
- 2008-07-21 KR KR1020080070865A patent/KR20090016383A/en not_active Application Discontinuation
- 2008-08-11 CN CN2008102102790A patent/CN101365093B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JP2009065630A (en) | 2009-03-26 |
| JP5040687B2 (en) | 2012-10-03 |
| KR20090016383A (en) | 2009-02-13 |
| CN101365093A (en) | 2009-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3927557B2 (en) | Video signal deinterlacing method and deinterlacing apparatus | |
| US8830403B1 (en) | Image processing device and image processing method | |
| CN101674448B (en) | Video signal processing device, video signal processing method | |
| US8290310B2 (en) | Image processing apparatus and method, program, and recording medium | |
| EP2227902B1 (en) | Interpolation frame generation apparatus, interpolation frame generation method, and broadcast receiving apparatus | |
| EP4287633A1 (en) | Video frame interpolation method and apparatus, and electronic device | |
| JP5669523B2 (en) | Frame interpolation apparatus and method, program, and recording medium | |
| JPWO2009107487A1 (en) | Motion blur detection device and method, image processing device, and image display device | |
| CN1326395C (en) | Data processing method,data processing device | |
| CN112866795B (en) | Electronic apparatus and control method thereof | |
| CN101365093B (en) | Image processing apparatus and method | |
| CN101651813A (en) | Method of driving display apparatus and driving circuit for display apparatus using the same | |
| US20100039517A1 (en) | Film cadence detection | |
| CN101616291B (en) | Image processing apparatus and method and program | |
| US8331676B2 (en) | Image processing apparatus and method, and program | |
| JP5015089B2 (en) | Frame rate conversion device, frame rate conversion method, television receiver, frame rate conversion program, and recording medium recording the program | |
| KR20110048252A (en) | Method and apparatus for image conversion based on sharing of motion vector | |
| JP5123643B2 (en) | Video processing device | |
| JP2008035436A (en) | Pull-down sequence detecting apparatus and method | |
| CN113452933B (en) | Video signal conversion method and display device | |
| CN115334334B (en) | Video frame inserting method and device | |
| CN103581639A (en) | Method and system for processing 3D images | |
| JP2009124261A5 (en) | ||
| JP5549431B2 (en) | Server apparatus, moving vector detection method and program | |
| CN117762247A (en) | Image display self-adaptive adjusting method, terminal equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120208 Termination date: 20130811 |