WO2004002148A1 - Dispositif de determination de vecteur de mouvement, procede de determination, dispositif de compensation de mouvement et procede de compensation de mouvement - Google Patents

Dispositif de determination de vecteur de mouvement, procede de determination, dispositif de compensation de mouvement et procede de compensation de mouvement Download PDF

Info

Publication number
WO2004002148A1
WO2004002148A1 PCT/JP2003/007438 JP0307438W WO2004002148A1 WO 2004002148 A1 WO2004002148 A1 WO 2004002148A1 JP 0307438 W JP0307438 W JP 0307438W WO 2004002148 A1 WO2004002148 A1 WO 2004002148A1
Authority
WO
WIPO (PCT)
Prior art keywords
pixel
field
vector
motion
block
Prior art date
Application number
PCT/JP2003/007438
Other languages
English (en)
Japanese (ja)
Inventor
Takaya Hoshino
Toshio Sarugaku
Makoto Kondo
Kazuhiko Nishibori
Koji Aoyama
Yukihiko Mogi
Original Assignee
Sony Corporation
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Sony Corporation filed Critical Sony Corporation
Publication of WO2004002148A1 publication Critical patent/WO2004002148A1/fr

Links

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
    • H04N7/0127—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level by changing the field or frame frequency of the incoming video signal, e.g. frame rate converter
    • H04N7/0132—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level by changing the field or frame frequency of the incoming video signal, e.g. frame rate converter the field or frame frequency of the incoming video signal being multiplied by a positive integer, e.g. for flicker reduction
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/144—Movement detection
    • H04N5/145—Movement estimation
    • 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
    • H04N7/0135—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level involving interpolation processes
    • H04N7/014—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level involving interpolation processes involving the use of motion vectors

Definitions

  • the present invention relates to a motion vector detection device and a detection method capable of reducing a circuit scale while reducing an influence on image quality, and a motion vector detection device and a detection method.
  • the present invention relates to a motion correction device and a motion correction method.
  • This interlaced scanning system is adopted as a television standard system in various countries around the world.
  • PAL Phase Alternation by Line
  • the field frequency is 50%.
  • [Hz] frame image is 25 frames / second
  • field image is 50 fields Z seconds.
  • the input image signal is subjected to processing such as internal processing in order to further suppress the interference of the plane frit, so that the field frequency is doubled from 50 Hz to 100 Hz.
  • the field frequency double speed method which converts the image signal into an image signal, has conventionally been adopted.
  • FIG. 1 shows a block diagram of a field double speed conversion circuit 7 to which the field frequency double system is applied.
  • This field double speed conversion circuit 7 has input terminals 81 and The CRT 83 and the horizontal / vertical deflection circuit 82 are integrated in a television receiver 8.
  • the field double speed conversion circuit 7 includes a double speed conversion unit 71 and a frame memory 72.
  • the double-speed conversion unit 71 writes, for example, an image signal of 50 fields / second of the PAL system input from the input terminal 81 to the frame memory 72.
  • the double-speed conversion section 71 reads out the image signal written in the frame memory 72 at twice the speed of writing. This makes it possible to double the frequency of the image signal of 50 fields / second and generate an image signal of 100 field seconds.
  • the double speed conversion section 71 outputs the double speed converted image signal to the CRT 83.
  • the CRT 83 displays the input image signal on the screen.
  • the horizontal and vertical deflection of the image signal in the CRT 83 is controlled based on a horizontal / vertical sawtooth wave generated by the horizontal / vertical deflection circuit 82 and having twice the frequency of the input image signal.
  • the image signal indicated by a white circle in FIG. 2A is an interlaced image signal of 50 fields Z seconds before double-speed conversion, and the image signal indicated by a black circle in FIG. This is a field / second interval image signal.
  • field f1 and field f2 are signals created from the same frame of the film, and similarly, field f3 and field f4 also form the same frame. I do. Since these image signals are interlaced image signals, pixel positions in the vertical direction differ between adjacent fields. For this reason, it is not possible to create one new field between each field while maintaining the interlacing property.
  • two new fields f′2 and f′1 are generated between the field f1 and the field f2.
  • No field is generated between field f2 and field f3
  • two new fields f'4 and f'3 are generated between field f3 and field f4. That is, one frame is formed by four fields and two frames.
  • the newly generated fields f'1, f'2, ... May be obtained as a median value of three pixels around each pixel using a median filter or the like.
  • the newly created fields: f'1, f'2, ... have the same contents as the fields fl, f2, ..., respectively.
  • the field double-speed conversion circuit 5 alternately arranges a part for newly generating two fields and a part for no generation between the fields of the image signal before the double-speed conversion, thereby obtaining the number of screens per unit time. Can be increased, and it is possible to suppress the above-described surface frit force disturbance.
  • FIG. 16 shows the relationship between each field and the image position when the image moves in the horizontal direction in the image signal after the telecine conversion.
  • the horizontal axis indicates the position of the image in the horizontal direction
  • the vertical axis indicates time.
  • the fields fl and f2 constitute the same frame, so that the image is displayed at the same position. This image moves horizontally (to the right in Fig. 3) when moving to field 3.
  • Field f4 is displayed at the same position as field f3 because it constitutes the same frame as field f3.
  • the fields f1, f'2, and f 'constituting the same frame are obtained.
  • the same image is displayed at the same position with 1, f2.
  • the same image is displayed at the same position in the fields f3, f'4, f'3, and f4 that form the same frame.
  • Figure 4A shows the relationship between each field and the image position when the image moves in the horizontal direction in the television signal before double-speed conversion (hereinafter referred to as TV signal).
  • TV signal double-speed conversion
  • the image signal of the TV signal shown in Fig. 4A is doubled by the field frequency double speed method.
  • the same image is displayed at the same position in the fields f 1 and f ′ 2 constituting the same frame.
  • the same image is displayed at the same position in the fields f ′ 1 and f 2 constituting the same frame.
  • the image is displayed at the same position up to the fields f1 to f2, while shifting to f2 to f3 If you want to move greatly in the horizontal direction.
  • the image signal obtained by converting the TV signal at double speed as shown in FIG. 4B, the image is displayed at the same position in the fields fl to f′2, but shifts to ⁇ ′2 to f′1 Move in the horizontal direction.
  • the output image signal regularly configures each field with a period of 1/1100 second, the time period during which the image operates is shorter than the time period during which the image is stationary. There is a problem in that when viewing a program via a TV, the motion of the image looks discontinuous.
  • the motion vector is obtained by dividing the screen into blocks of predetermined pixels based on, for example, a block matching method, and evaluating the similarity for each block.
  • the motion was corrected by shifting the pixel position for each block according to the obtained motion vector.
  • a reference field is divided into a plurality of reference blocks 101, and a block having the highest similarity to the reference block 101 in the reference field 80 is calculated as follows: It is detected from the search block 103 moving in the search range 104 in the reference field 90. The displacement between the detected search block 103 and the reference block 101, that is, the direction and magnitude of the movement is defined as the moving vector.
  • the above-described determination of the degree of approximation is performed by first taking the difference between each pixel value of the search block 103 and the corresponding pixel value of the reference block 101, and evaluating the difference indicated by the difference, for example, the sum of absolute differences. Ask for. Next, the above-described determination operation is performed for all search blocks 103, and the smallest evaluation value is obtained from the sum of evaluation values obtained, that is, the sum of absolute differences.
  • the search block 103 giving the minimum sum of absolute differences is a block showing the highest similarity to the reference block 101, and a motion vector is specified according to the pixel position of the block.
  • FIG. 6A is a diagram illustrating a case where motion compensation is performed on each pixel of the image signal subjected to the double speed conversion after the telecine conversion illustrated in FIG. 3A
  • FIG. 6B is a diagram illustrating a case where the TV signal illustrated in FIG. 4B is double speed converted.
  • FIG. 3 is a diagram illustrating a case where motion correction is performed for each pixel of an image signal.
  • the motion vector a 1 1 is obtained between the field f 1 and the field f 3, and in the vector direction of the obtained motion vector a 11 1 and within the range of the vector amount, the field f ′ 2 to field; each pixel of f2 is gradually shifted.
  • the motion vector b 11 is obtained between the field f 1 and the field f ′ 1, and the field f is set in the vector direction of the obtained motion vector b 11 1 and within the range of the vector amount. 'Shift 2 pixels.
  • a motion vector b 1 2 is obtained between the field f ′ 1 and the field f 3, and in the vector direction of the obtained motion vector b 1 2 and within the range of the vector amount, the field f Shift 2 pixels.
  • the above-described reference block 101 and search block 103 are simply obtained by dividing the screen into a set of a predetermined number of pixels, and in these blocks 101 and 103, pixels are spread without gaps. It is configured.
  • an absolute difference value is calculated for all pixels constituting each block 101, 103.
  • the image signal generated by telecine conversion and double speed conversion is composed of 4 fields per frame, so the motion vector obtained by comparing with the image signal composed of 2 fields per frame obtained by converting the TV signal to double speed Is large (see all in Fig. 6A, bll in head 6B, b12). For this reason, when motion-correcting an image signal after telecine conversion, the search range 104 is wider than that of a TV signal.
  • the search block 103 must be sequentially moved after setting, and the circuit scale is further increased due to the huge amount of calculation of the absolute difference value.
  • the present invention is to provide a novel motion vector detecting device and method, and a motion correcting device and a motion correcting method that can solve the problems of the conventional technology as described above. .
  • Another object of the present invention is to provide a motion vector detecting apparatus and a detecting method, and a motion correcting apparatus and a motion correcting method capable of reducing the amount of calculation of the absolute difference value while reducing the influence on the image quality. Is to provide.
  • the motion vector detection device and detection method according to the present invention proposed to achieve the above-described object include a reference pixel in a reference field one frame or two frames before an input reference field.
  • the sum of absolute differences of the pixel values is sequentially calculated between the reference block having the origin as the origin and the search block in which the reference field is moved at least every other pixel vertically or at least every other pixel horizontally,
  • the motion vector between the pixel position of the search block and the pixel position of the reference pixel that minimizes the calculated sum of absolute differences, the amount of calculation while maintaining the required accuracy of the motion vector And reduce the circuit size.
  • the motion vector detection device is a motion vector detection device that detects a motion vector by a block matching method for each pixel of a reference field one frame or two frames before an input reference field.
  • a reference block starting from the reference pixel in the reference field is cut out, and a search block in the reference field is placed at least every other vertical pixel and / or at least one horizontal picture.
  • Block extracting means for moving and cutting out every other; vector calculating means for sequentially calculating the sum of absolute differences of pixel values between a reference block and a search block cut out by the block extracting means; A vector specifying unit that specifies a motion vector between the pixel position of the search block and the pixel position of the reference pixel in which the sum of absolute differences calculated by the vector calculating unit is a minimum.
  • the motion vector detection method is a motion vector detection method for detecting a motion vector by a block matching method for each pixel of a reference field one frame or two frames before an input reference field.
  • a reference block having the reference pixel in the field as the origin is cut out, and the search block in the reference field is cut out by moving at least every other pixel vertically and / or at least every other pixel horizontally, and the cut out reference block and
  • the sum of the absolute values of the pixel values is sequentially calculated between the search block and, and the motion vector is specified between the pixel position of the search block and the pixel position of the reference pixel where the calculated sum of the absolute differences is minimum-
  • the motion compensation device and method according to the present invention may be configured such that one frame Or, the pixel value between the reference block whose origin is the reference pixel in the reference field two frames before, and the search block in which the reference field is moved at least every other pixel vertically and / or at least every other horizontal pixel The motion vector is detected by sequentially calculating the sum
  • the motion compensation device is a motion compensation device that performs motion compensation for each pixel of a reference field two frames before a reference field to be input.
  • the sum of the absolute differences of the pixel values is sequentially calculated between the block and the search block in which the reference field is moved at least every other pixel or at least every other pixel, and the calculated sum of the absolute differences is calculated.
  • Vector specifying means for specifying the motion vector between the pixel position of the search block and the pixel position of the reference pixel, which is the minimum, and the amount of shift in the vector direction of the motion vector specified by the vector specifying means
  • Pixel shift means for shifting the position of the reference pixel in the reference field so that the value falls within the range of the vector amount.
  • Another motion compensating device uses one frame from an input reference field.
  • a motion compensator for motion compensating each pixel of a previous reference field, a reference block whose origin is the reference pixel in the reference field and a reference field at least every other vertical pixel or Z and at least horizontal
  • the sum of the absolute values of the pixel values is sequentially calculated with the search block shifted every other pixel, and the pixel position of the search block and the pixel position of the reference pixel at which the calculated sum of the absolute differences is minimized
  • a vector specifying means for specifying a motion vector between the reference field and a reference field such that a shift amount in the vector direction of the motion vector specified by the vector specifying means is within the range of the vector amount.
  • a pixel shifting means for shifting the position of the reference pixel.
  • Still another motion compensating apparatus is a motion compensating apparatus which performs motion compensation for each pixel of a reference field one frame before or two frames before an input reference field.
  • As bets amount is in the range of a vector quantity, the position of the reference pixels definitive based field and a pixel shifting means for shifting Bok.
  • a motion compensation method is a motion compensation method for performing motion compensation for each pixel of a reference field two frames before an input reference field, comprising: a reference block having a reference pixel in a reference field as an origin; The sum of the absolute differences of the pixel values is sequentially calculated between the search block in which the field is moved at least every other pixel or at least every Z and at least every other horizontal pixel, and the calculated sum of the absolute differences is minimized.
  • the motion vector is specified between the pixel position of the search block to be searched and the pixel position of the reference pixel.
  • the reference is set so that the shift amount in the vector direction of the specified motion vector is within the range of the vector amount. Shift the position of the reference pixel in the field.
  • Another motion compensation method is a motion compensation method for performing motion compensation for each pixel of a reference field one frame before a reference field to be inputted.
  • the sum of the absolute differences of the pixel values is sequentially calculated between the reference block whose origin is the reference pixel from the field and the search block in which the reference field is moved at least every other pixel vertically or at least every other pixel horizontally,
  • a motion vector is specified between the pixel position of the search block and the pixel position of the reference pixel, at which the calculated sum of absolute differences is minimized, and the shift amount in the vector direction of the specified motion vector is the vector amount.
  • the position of the reference pixel in the reference field is shifted so as to fall within the range.
  • Still another motion compensation method is a motion compensation method for performing motion compensation for each pixel of a reference field one frame before or two frames before an input reference field.
  • the reference field of the previous or two frames before is switched, and the reference block having the reference pixel as the origin in the switched reference field and the reference field are moved at least every other pixel vertically and / or at least every other pixel horizontally.
  • the sum of the absolute values of the pixel values is sequentially calculated with the search block, and the motion vector between the pixel position of the search block where the calculated sum of the absolute differences is the minimum and the pixel position of the reference pixel is calculated.
  • the shift amount is within the range of the vector amount in the vector direction of the specified motion vector. Shifting the position of the original.
  • FIG. 1 is a block diagram of a field double speed conversion circuit to which the field frequency double speed method is applied.
  • 2A and 2B are diagrams showing the relationship between each field and the pixel position before and after the double-speed conversion.
  • FIGS. 3A and 3B are diagrams showing the relationship between each field and the image position when the image moves in the horizontal direction when the telecine conversion signal is input ⁇ FIG. 4A and FIG. B means that when a TV signal is input, the image FIG. 4 is a diagram illustrating a relationship between each field and an image position when moving.
  • FIG. 5 is a diagram for explaining a conventional block matching method.
  • FIG. 6A and FIG. 6B are diagrams showing an example of obtaining a motion vector for a telecine conversion signal and a TV signal.
  • FIG. 7 is a block diagram showing a motion correction device according to the present invention.
  • 8A and 8B are diagrams showing the relationship between each field and the pixel position before and after the double speed conversion in the field double speed conversion circuit.
  • FIG. 9 is a diagram showing the relationship between each field and the image position when the image moves in the horizontal direction in the TV signal.
  • FIGS. 10A and 10B are diagrams for explaining an operation when a TV signal is input in the motion vector detection unit.
  • FIG. 11 is a diagram showing an example in which the pixel position is shifted with respect to the input TV signal in the image shift unit.
  • Fig. 12 shows an example where the shift amount of the first field is set to 0 and the shift amount of the second field is set to 1Z2 times the vector amount of the detected motion vector in the image shift unit. It is.
  • FIG. 13 is a diagram showing the relationship between each field and the image position when the image moves in the horizontal direction in the telecine conversion signal.
  • FIGS. 14A and 14B are diagrams for explaining the operation when a telecine conversion signal is input to the motion vector detection unit.
  • FIG. 15 is a diagram illustrating an example in which the image shift unit shifts the pixel position of the input telecine conversion signal.
  • FIG. 16 is a diagram showing an example in which the pixel position is shifted in the image shift unit, so that the image moves smoothly.
  • FIG. 17 is a diagram showing an example in which the shift amount of the first field is set to 0 in the image shift unit, and thereafter, the amount of the motion vector is increased by ⁇ times.
  • FIG. 18A and FIG. 18B are diagrams showing an example of detecting a motion vector based on each block formed by thinning out every vertical and horizontal pixels.
  • Figures 19A and 19B show at least every other pixel vertically and / or at least FIG. 9 is a diagram showing an example in which a search block 53 is moved every other pixel.
  • BEST MODE FOR CARRYING OUT THE INVENTION a motion correction device and method thereof, and a motion vector detection device and detection method to which the present invention is applied will be described in detail with reference to the drawings.
  • the motion compensation apparatus 1 is built in a television receiver of, for example, a PAL (Phase Alternation by Line) method, and as shown in FIG. 7, a first image memory 11 and a second An image memory 12, a switch 13, a motion vector detection unit 14, and an image shift unit 15 are provided.
  • PAL Phase Alternation by Line
  • the first image memory 11 stores an interlaced image signal (hereinafter referred to as a TV signal) in which one frame generated by converting a television signal at double speed into two fields, or an image obtained by telecine conversion.
  • An in-lace image signal (hereinafter referred to as a telecine conversion signal) in which one frame generated by double-speed conversion is composed of four fields is sequentially supplied.
  • the first image memory 11 stores the supplied image signal for one frame in each field unit. That is, the image signal output from the first image memory 11 is one frame before the image signal supplied to the first image memory 11.
  • the second image memory 12 has the same internal configuration as the first image memory 11, and stores the image signal supplied from the first image memory 11 for each field in units of one frame. I do. That is, the image signal output from the second image memory 12 is one frame earlier than the image signal supplied to the second image memory 12 and is supplied to the first image memory 11 Two frames before the image signal.
  • the image signal stored in the second image memory 12 is supplied to an image shift unit 15 and a motion vector detection unit 14.
  • the image signal transmitted from the second image memory 12 to the motion vector detection unit 14 is referred to as a reference field 30.
  • the switch 13 sequentially switches between the image signal supplied to the first image memory 11 and the image signal output from the first image memory 11, and transmits this to the motion vector detection unit 14.
  • this switch 13 sends a signal to the motion vector detector 14.
  • the image signal to be referred to is referred to as a reference field 40.
  • the switch 13 When a TV signal is input, the switch 13 is switched to the “Normal” (N) side in FIG. 7, and the reference field 40 transmitted to the motion vector detector 14 is a reference field. One frame after 30.
  • the switch section 13 is switched to the “film” (F) side in FIG. 7, and the reference field 40 transmitted to the motion vector detection section 14 is Two frames after the reference field 30.
  • the motion vector detector 14 detects the motion vector for each block or pixel based on the supplied reference field 30 and reference field 40, and detects the detected motion vector or shift based on the motion vector.
  • the information about the amount and the shift direction is output to the image shift unit 15.
  • the image shift unit 15 receives the motion vector detected by the motion vector detection unit 14 or information such as a shift amount based on the motion vector for each block or each pixel.
  • the image shift unit 15 shifts each pixel position or each block in the image signal supplied from the second image memory 12 in the vector direction within the range of the vector amount of the received motion vector.
  • the image shift unit 15 outputs the image signal shifted in the vector direction to CRT2.
  • the CRT 2 displays the input image signal on the screen.
  • the horizontal and vertical deflection of the image signal in the CRT2 may be controlled based on a horizontal / vertical deflection circuit (not shown).
  • the motion correction device 1 may integrate a field double-speed conversion circuit 3 that double-converts the field frequency of an image signal.
  • the field double speed conversion circuit 3 is integrated to improve the resolution and prevent the disturbance of the plane flit force. For example, by performing processing such as interpolation in the PAL method, the field frequency is reduced to 50H.
  • the image signal of z is converted into an image signal of 100 Hz twice.
  • the field frequency conversion circuit 3 includes an input terminal 31 connected to a television receiver, a double speed conversion unit 32, and a frame memory 33.
  • the double-speed converter 32 writes the telecine-converted image signal input from the television receiver via the input terminal 31 to the frame memory 33.
  • the double-speed conversion unit 32 converts the image signal written in the frame memory 33 into a signal twice as large as that in the writing. Read at speed. As a result, for example, the frequency of an image signal of 50 fields / second in the PAL system can be doubled, and an image signal of 100 fields / second can be generated.
  • the double-speed conversion unit 32 supplies the double-speed-converted image signal to the motion correction device 1.
  • FIG. 8A and 8B show the relationship between each field and the pixel position before and after the double speed conversion in the field double speed conversion circuit 3.
  • FIG. Here, the horizontal axis indicates time, and the vertical axis indicates the vertical position of the pixel.
  • the image signal before the double-speed conversion is a PAL-type interlace image signal of 50 fields / sec. As shown in FIG. 8A, one frame is formed by two fields. On the other hand, since the image signal after double-speed conversion is an in-race image signal of 100 fields // second, as shown in FIG. 8B, a new signal between field t1 and field t2 is newly added. Generate two fields t 2, t '1. Then, no field is generated between the field t2 and the field t3, and two new fields t'4 and t'3 are generated between the field t3 and the field t4. That is, in the case of an image signal, for example, a telecine conversion signal, one frame is formed in four fields.
  • Each pixel value may be obtained as an intermediate value of three pixels around each pixel.
  • the newly generated fields t'1, t'2,... Have the same contents as the fields tl, t2,.
  • one frame is formed in four fields, and the resolution can be improved by increasing the number of screens per unit time, and it is possible to suppress surface flicker interference Becomes
  • a field double speed conversion circuit 3 sequentially supplies a TV signal in which one frame generated by performing double speed conversion of a television signal is composed of two fields.
  • Figure 9 shows the relationship between each field and the image position when the image moves in the horizontal direction of the TV signal.
  • the horizontal axis represents the position of the image in the horizontal direction
  • the vertical axis represents time.
  • Field frequency double speed method The same image is displayed at the same position in the fields t 1 and t ′ 2 constituting the same frame in the TV signal that has been double-speed converted by the formula.
  • the same image is displayed at the same position in the fields t'l and t2 constituting the same frame.
  • Switch 13 receives the TV signal and switches to the “Normal” (N) side in FIG. 7, and the reference field 30 and the reference field 40 one frame after the reference field 30 move. It is supplied to the vector detector 14.
  • the motion vector detection unit 14 detects a motion vector between the reference field 30 and the reference field 40 for each pixel or each block based on, for example, a block matching method.
  • the motion vector detector 14 divides the reference field 30 into a reference block 51 composed of m ⁇ n pixels centered on a reference pixel 58 as shown in FIG.
  • the block having the highest degree of approximation of 1 is detected from the search block 53 moving in the search range 54 in the reference field 40 shown in FIG. 10B.
  • the search block 53 is composed of m ⁇ n pixels having the same size as the reference block 51 centered on the center pixel 59, and the search range 54 in which the search block 53 moves is composed of M ⁇ N pixels.
  • the motion vector detection unit 14 specifies a motion vector between the reference block 51 and the center pixel 59 of the search block 53 having the highest similarity.
  • the above-described determination of the degree of approximation is performed by first taking the difference between each pixel value of the search block 53 and the corresponding pixel value of the reference block 51, and evaluating the difference indicated by the difference, for example, the sum of absolute differences. Ask for. Next, the above-described determination operation is performed for all the search blocks 53, and the evaluation value obtained for each of the search blocks 53, that is, the smallest one from the sum of absolute differences is obtained.
  • the search block 53 that gives the minimum sum of absolute differences is a block showing the highest similarity to the reference block 51, and a moving vector is specified according to the pixel position of the block.
  • the search block 53 moves so that the center pixel 59 scans all the pixels constituting the search range 54 shown in FIG. 10B.
  • the search block 53 moves the search range 54 so as to be shifted by one pixel.
  • the required accuracy of the motion vector can be improved.
  • the vector direction of the motion vector is the reference field 3
  • the direction becomes rightward in the figure.
  • the vector amount becomes C.
  • the reference field is t3
  • the reference field is t'1
  • the vector amount of the motion vector is D.
  • the image shift unit 15 shifts the position of the detection pixel in the vector direction of the image signal supplied from the first image memory 12 based on the context of each field.
  • the first field to be entered is the first field
  • the second field to be entered after the first field is the second field.
  • the shift amount is sequentially increased so as to be equal to the first field and the second field.
  • the image When the image is shifted as described above, it can be indicated by the dotted line in FIG. 11, but the image is shifted from the second field to the first field as compared with before the image is shifted. This makes it easier to move the image without making large movements.
  • the field double speed conversion circuit 3 By integrating the field double speed conversion circuit 3 into the motion compensation device 1 having such a configuration and incorporating it in the television receiver, it is possible to eliminate the discontinuity of motion in the image obtained by converting the TV signal at double speed. it can. In other words, the image quality can be synergistically improved by improving the resolution by the field double speed conversion circuit 3 and smoothing the movement of each image in which the surface flicker interference is suppressed.
  • the shift amount of the image in the image shift unit 15 is, for example, as shown in FIG. 12, the shift amount of the first field is set to 0, and the shift amount of the motion vector detected by It is also possible to make the amount of vector 1 to 2 times.
  • the shift amount of the first field is set to 0, and when the vector amount is C, the second field is shifted by C X 1/2. If the vector quantity is D, the second field is shifted by D X 1 Z 2. Since one frame is composed of two fields, it is possible to increase the shift amount linearly with time by increasing the vector amount by 1/2 times, making the image movement even smoother. can do.
  • the motion compensation device 1 sequentially receives from the field double speed conversion circuit 3 a telecine conversion signal in which one frame generated by performing double speed conversion of the telecine-converted image is composed of four fields.
  • Figure 13 shows the relationship between each field and the image position when the image moves in the horizontal direction of the telecine conversion signal.
  • the horizontal axis represents the position of the image in the horizontal direction
  • the vertical axis represents time.
  • the telecine conversion signal double-speed converted by the field frequency conversion method is supplied to the first image memory 11 at fixed time intervals in the order of fields tl, t'2, t'1, and t2, and all the images are It is displayed at the same position.
  • the image shifts in the horizontal direction (rightward), and is supplied to the first image memory 11 in the order of the fields t3, t'4, t'3, and t4.
  • the switch 13 receives the telecine conversion signal and switches to the “film” (F) side in FIG. 7, and sets the reference field 30 and the reference field two frames after the reference field 30. 40 is supplied to the motion vector detection unit 14.
  • the motion vector detection unit 14 detects a motion vector between the reference field 30 and the reference field 40 for each pixel or each block based on, for example, a block matching method.
  • the motion vector detection unit 14 uses a reference field 30 as shown in FIG.
  • the reference block 51 is divided into a reference block 51 composed of m ⁇ n pixels centered on the reference pixel 58, and a block having the highest similarity to the reference block 51 is referred to as a reference field 40 shown in FIG. 14B.
  • a reference field 40 shown in FIG. 14B.
  • the search block 53 is composed of m ⁇ n pixels of the same size as the reference block 51, centered on the center pixel 59, and the search range 54 in which the search block 53 moves is 2 MX 2 N It is composed of pixels and is twice as long and wide in area and four times as large as the TV signal search range 54.
  • the search range of the telecine conversion signal needs to be set to twice as large in the vertical and horizontal directions as described above.
  • the search range is not limited to twice the vertical and horizontal lengths of the TV signal search range 54, and may be configured with any size.
  • the motion vector detection unit 14 specifies a motion vector between the reference block 51 and the center pixel 59 of the search block 53 having the highest similarity.
  • the search block 53 moves such that the center pixel 59 scans only the colored pixels every other vertical and horizontal pixels constituting the search range 54 shown in FIG. 14B. In other words, the search block 53 moves in such a way that the search range 54 is shifted every other pixel vertically and horizontally.
  • the search range 54 consists of 2 MX 2 N pixels in the case of a telecine conversion signal, which is twice as long and wide (the area is 4 times) as compared to the search range 54 for a TV signal. By thinning out every other pixel in the vertical and horizontal directions, the amount of calculation of the absolute difference can be made equal to the search range 54 composed of the MXN pixels of the TV signal.
  • a true motion vector when a telecine conversion signal is input, a true motion vector can be obtained by making the search range 54 four times as large as the search range of the TV signal. By scanning 53 every other pixel in the vertical and horizontal directions, the calculation amount of the absolute difference can be reduced, and the circuit scale can be reduced.
  • the motion compensation is performed by sequentially switching the image signal after conversion and the TV signal.
  • the circuit scale can be unified by using the same amount of calculation as when a TV signal is input, especially for a telecine conversion signal in which the calculation amount of the sum of absolute differences is enormous. Therefore, the specifications of the television receiver can be satisfied.
  • the vector direction of the motion vector obtained by the motion vector detection unit 14 is rightward with respect to the reference field 30, and when the reference field is t 3.
  • the vector quantity is A.
  • the reference field is t5
  • the reference field is t3
  • the vector amount of the motion vector is B.
  • the image shift unit 15 shifts the position of the detection pixel in the vector direction in the image signal supplied from the second image memory 12 based on the context of each field.
  • the first field to be entered is the first field
  • the next field to be entered after the first field is the second field
  • the shift amount of each field is as follows: the first field ⁇ the second field followed by the third field ⁇ the fourth field, as shown in Figure 15 And so on.
  • the image moves significantly when moving from the fourth field to the first field as compared to before the image is shifted, as shown in FIG. Without moving the image, the motion of the image can be smoothed.
  • the circuit scale can be reduced to the TV signal even when the telecine conversion signal is input. It is possible to eliminate the inconsistency of movement while keeping it at the same level. That is, the resolution is improved by the field double-speed conversion circuit 3, and the movement of each image in which the surface fringe disturbance is suppressed is further smoothed, so that the image quality can be synergistically improved.
  • This motion compensator 1 can be used not only when implemented alone, A remarkable effect can be obtained by implementing the method integrally with the circuit 3. Even a television receiver having a field double-speed conversion circuit already integrated therein can easily be upgraded by incorporating the motion compensation device 1 later.
  • the image shift amount in the image shift unit 15 is, for example, as shown in FIG. 17, the shift amount of the first field is set to 0, and the shift amount of the second field is detected. It is possible to increase the amount by 1/4 times the vector amount of the vector, and then increase it by 1/4 times.
  • the shift amount of the first field is set to 0, and if the vector amount is A, the second field is shifted by AX 1 Z 4 and the third field is shifted by AX 2 Z 4
  • the shift amount can be linearly increased with time, and the motion of the image can be further smoothed.
  • the motion vector required may have an error of one pixel ( however, this image shift unit 15 Since the pixels are shifted by the number obtained by dividing the obtained motion vector by 4, the error for one pixel is also divided by 4, and the visual inconvenience caused by such pixels is almost eliminated.
  • image quality is significantly degraded due to discontinuous motion when no motion correction is performed, but by performing motion correction using the motion correction device 1 according to the present invention. Since the image quality can be significantly improved, even if the required accuracy of the motion vector is slightly lower, there is almost no visual effect.
  • a low-pass filter (not shown) can be applied in advance before the motion vector detecting section 14 specifies a motion vector.
  • LPF low-pass filter
  • a motion vector may be detected by the following method.
  • the motion vector detector 14 divides the reference field 30 into a reference block 61 composed of 2 mx 2 n pixels centered on the reference pixel 58 as shown in FIG. the block exhibiting the highest degree of approximation to the block 61, FIG. 1 8 c the reference proc 6 1 detected from the search block 6 3 moving the search range 6 4 in the reference field 4 0 shown in B, the above-described Compared to the reference block 51 composed of m X n pixels, the height and width are twice as large and the area is four times.
  • the search block 63 is composed of 2 mx 2 n pixels of the same size as the reference block 61 with the center pixel 59 at the center, and is thinned out every other pixel in the vertical and horizontal directions.
  • the search range 64 in which the search block 63 moves is composed of 2 MX 2 N pixels, and is twice as long and wide in area and four times as large as the search range 54 for TV signals.
  • the motion vector detection unit 14 specifies a motion vector between the reference block 51 and the center pixel 59 of the search block 63 exhibiting the highest degree of approximation.
  • the size of each of the blocks 61 and 63 and the search range 64 may be any size without being limited to the above.
  • the search block 53 moves such that the center pixel 59 scans only the colored pixels every other vertical and horizontal pixels constituting the search range 54 shown in FIG. 14B.
  • the search block 63 moves so that the search range 64 shown in FIG. 18B is shifted every other pixel vertically and horizontally.
  • the search range 6 4 is composed of 2 MX 2 N pixels in the case of a telecine conversion signal, and is twice as long and wide (the area is 4 times) as compared with the search range 6 4 for a TV signal. By performing the search by thinning out every other pixel in the vertical and horizontal directions, the calculation amount of the absolute difference value can be made equal to the search range 54 composed of the MXN pixels of the TV signal.
  • the motion vector detection unit 14 sets the center pixel 59 on the search ranges 54 and 64 as shown in FIGS. 14B and 18B.
  • the search is performed so as to scan over the colored pixels, in other words, the search blocks 53 and 63 are shifted every other pixel in the vertical and horizontal directions, but this is not a limitation.
  • a search may be performed by shifting only the horizontal at every other pixel, and a search may be performed without shifting the vertical at every other pixel. Similarly, shift only vertical by every other pixel The search may be performed without changing the horizontal position every other pixel.
  • the interval at which the pixels are shifted is not limited to one pixel, but it is sufficient if at least one pixel is shifted, for example, every two pixels or every third pixel.
  • FIG. 19B shows an example in which a search is performed by shifting the vertical by every other pixel and the horizontal by every three pixels. That is, in the present invention, any search method may be used as long as the search is performed by shifting the search block at least every other pixel or at least every other horizontal pixel. This makes it possible to reduce the circuit scale while making the search ranges 54 and 64 wide.
  • the reference block 61 and the search block 63 shown in Fig. 18A are not limited to the case where they are thinned out every other vertical and horizontal pixels, but also at least every other vertical pixels and / or at least every other horizontal pixels. It is enough if it is thinned out. Therefore, it is a matter of course that block matching may be performed using the reference block 61 and the search block 63 thinned out, for example, every two or three pixels.
  • the motion vector detection unit 14 searches not only when a telecine conversion signal is input but also every other vertical and horizontal pixels as shown in FIGS. 14B and 18B described above when a TV signal is input.
  • the search may be performed by moving block 53.
  • the search may be performed by moving the search block 53 at least every other vertical pixel and / or at least every other horizontal pixel as shown in FIG. 19B described above.
  • LPF not-shown single pass filter
  • the motion compensator 1 is not limited to the above-described embodiment.
  • the switch 13 can be always switched to the “normal” (N) side. .
  • the reference field 40 transmitted to the motion vector detection unit 14 is always one frame later than the reference field 30. Therefore, the motion compensator 1 employing such a configuration performs motion compensation only for the TV signal.
  • a motion compensation circuit can be configured.
  • the motion vector detection unit 14 performs a search block 5 every other pixel in the vertical and horizontal directions as shown in FIG. 14B and FIG. The search is performed by scanning 3. As a result, the motion vector can be And the circuit scale can be reduced.
  • the motion compensator 1 is not limited to the above-described embodiment.
  • the switch 13 can be always switched to the “film” (F) side. is there.
  • the reference field 40 transmitted to the motion vector detecting unit 14 is always two frames later than the reference field 30.
  • a motion compensation circuit that can perform motion compensation only is configured.
  • the amount of operation of the absolute difference value can be reduced, and the circuit scale can be reduced.
  • the present invention is not limited to the case where the present invention is applied to a television receiver based on the PAL system.
  • the present invention is applied to the NTSC (National TV System Co. i.e.
  • the present invention is also applicable to a television receiver to which an interlaced image signal of field seconds (30 frames / second) is input. It is also applicable to television receivers using the SECAM system.
  • the present invention can be incorporated not only into a television receiver but also into a signal converter connected to the television receiver.
  • the present invention can be applied to a case where an image signal transmitted on an Internet network is displayed on a PC or the like, or a case where a media image format is converted.
  • the present invention has been described in the form of being realized by hardware such as a circuit, it is needless to say that the present invention can also be realized by software on a processor.
  • the motion vector detecting apparatus and method according to the present invention can be used in the reference field one frame or two frames before the input reference field.
  • the absolute value of the difference in pixel value between the reference block whose origin is the reference pixel of interest and the search block that moves within the search range of the reference field at least every other vertical pixel or at least every Z and every horizontal pixel Since the motion vector is specified between the pixel position of the search block and the pixel position of the reference pixel where the calculated sum of absolute differences is minimized, the accuracy of the motion vector to be obtained is maintained.
  • the amount of calculation can be reduced, and the circuit scale can be reduced.
  • the motion compensation device and method according to the present invention provide a reference block having an origin at a reference pixel in a reference field one frame or two frames before an input reference field, and at least a search range in the reference field.
  • the motion vector is detected by sequentially calculating the absolute sum of the difference between the pixel value and the search block that moves every other vertical pixel or Z and at least every other horizontal pixel, and the motion vector is detected based on the motion vector. Since the position of the reference pixel in the reference field is shifted, the discontinuity of motion can be eliminated while reducing the amount of calculation and the circuit scale.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Television Systems (AREA)

Abstract

L'invention concerne un dispositif de détermination de vecteur de mouvement servant à déterminer un vecteur de mouvement par mise en correspondance de blocs en fonction de chaque pixel de la trame de base par une ou deux images avant la trame de référence d'entrée. Une somme de valeurs absolues de différence de valeurs de pixel est successivement calculée entre un bloc de base (51), ayant comme origine un pixel de référence (58) dans une trame de base (30) par une ou deux images avant une trame de référence d'entrée (40), et un bloc de recherche (53) dans lequel la trame de référence (40) est décalée au moins un pixel sur deux dans le sens longitudinal et/ou un pixel sur deux dans le sens latéral, un vecteur de mouvement entre la position de pixel du bloc de recherche (53) et la position de pixel du pixel de base (58) étant ainsi déterminé là où la somme des valeurs absolues de différence est minimale.
PCT/JP2003/007438 2002-06-19 2003-06-11 Dispositif de determination de vecteur de mouvement, procede de determination, dispositif de compensation de mouvement et procede de compensation de mouvement WO2004002148A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2002-179049 2002-06-19
JP2002179049A JP2004023673A (ja) 2002-06-19 2002-06-19 動きベクトル検出装置及び方法、動き補正装置及び方法

Publications (1)

Publication Number Publication Date
WO2004002148A1 true WO2004002148A1 (fr) 2003-12-31

Family

ID=29996550

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2003/007438 WO2004002148A1 (fr) 2002-06-19 2003-06-11 Dispositif de determination de vecteur de mouvement, procede de determination, dispositif de compensation de mouvement et procede de compensation de mouvement

Country Status (2)

Country Link
JP (1) JP2004023673A (fr)
WO (1) WO2004002148A1 (fr)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1592258B1 (fr) * 2004-04-30 2011-01-12 Panasonic Corporation Estimation de mouvement utilisant des vecteurs de reffinement spatiaux adaptatifs
WO2008136116A1 (fr) * 2007-04-26 2008-11-13 Pioneer Corporation Contrôleur de génération de trame d'interpolation, convertisseur de taux de trame, dispositif d'affichage, procédé pour commander la génération d'une trame d'interpolation, programme pour celui-ci, et support d'enregistrement stockant le programme
JP2009060401A (ja) 2007-08-31 2009-03-19 Sony Corp 画像表示装置および画像処理装置
JP4513913B2 (ja) * 2008-08-07 2010-07-28 ソニー株式会社 画像信号処理装置および方法
WO2010046990A1 (fr) * 2008-10-23 2010-04-29 パイオニア株式会社 Dispositif de génération de trame d'interpolation, dispositif de conversion de taux de trame, dispositif d'affichage, procédé de génération de trame d'interpolation, programme associé, et support d'enregistrement sur lequel son programme est enregistré
WO2010046989A1 (fr) * 2008-10-23 2010-04-29 パイオニア株式会社 Dispositif de conversion de taux de trame, dispositif de traitement d’images, dispositif d’affichage, procédé de conversion de taux de trame, programme associé, et support d'enregistrement sur lequel le programme est enregistré

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05137122A (ja) * 1991-11-14 1993-06-01 Hitachi Ltd 映像信号の信号処理方法及び回路
JPH05260460A (ja) * 1992-03-09 1993-10-08 Matsushita Electric Ind Co Ltd 動き検出装置
JPH06500910A (ja) * 1991-06-27 1994-01-27 トムソン　コンシュメ　エレクトロニクス 動画像のシーケンスのタイミングの変換方法
JPH07222176A (ja) * 1994-02-07 1995-08-18 Graphics Commun Lab:Kk 動画像圧縮におけるブロック・マッチング方法と装置
JPH0884293A (ja) * 1994-03-30 1996-03-26 Thomson Consumer Electron Sa 変換アーチファクトの減少方法及び装置
JPH0898186A (ja) * 1994-09-22 1996-04-12 Victor Co Of Japan Ltd 動きベクトル検出装置
JPH08223578A (ja) * 1995-02-13 1996-08-30 Nippon Telegr & Teleph Corp <Ntt> 動きベクトル探索方法および装置
JPH0918882A (ja) * 1995-07-03 1997-01-17 Matsushita Electric Ind Co Ltd 動きベクトル検索方法及びその装置
JPH09162747A (ja) * 1995-12-14 1997-06-20 Matsushita Electric Ind Co Ltd 動きベクトル検出装置およびその方法
JPH10501953A (ja) * 1995-04-11 1998-02-17 フィリップス エレクトロニクス ネムローゼ フェンノートシャップ 動き補償されたフィールドレート変換
JP2000023163A (ja) * 1998-06-30 2000-01-21 Sharp Corp 動きベクトル検出装置
JP2000102015A (ja) * 1998-09-22 2000-04-07 Sharp Corp 動きベクトル検出装置及び動きベクトル検出方法

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH06500910A (ja) * 1991-06-27 1994-01-27 トムソン　コンシュメ　エレクトロニクス 動画像のシーケンスのタイミングの変換方法
JPH05137122A (ja) * 1991-11-14 1993-06-01 Hitachi Ltd 映像信号の信号処理方法及び回路
JPH05260460A (ja) * 1992-03-09 1993-10-08 Matsushita Electric Ind Co Ltd 動き検出装置
JPH07222176A (ja) * 1994-02-07 1995-08-18 Graphics Commun Lab:Kk 動画像圧縮におけるブロック・マッチング方法と装置
JPH0884293A (ja) * 1994-03-30 1996-03-26 Thomson Consumer Electron Sa 変換アーチファクトの減少方法及び装置
JPH0898186A (ja) * 1994-09-22 1996-04-12 Victor Co Of Japan Ltd 動きベクトル検出装置
JPH08223578A (ja) * 1995-02-13 1996-08-30 Nippon Telegr & Teleph Corp <Ntt> 動きベクトル探索方法および装置
JPH10501953A (ja) * 1995-04-11 1998-02-17 フィリップス エレクトロニクス ネムローゼ フェンノートシャップ 動き補償されたフィールドレート変換
JPH0918882A (ja) * 1995-07-03 1997-01-17 Matsushita Electric Ind Co Ltd 動きベクトル検索方法及びその装置
JPH09162747A (ja) * 1995-12-14 1997-06-20 Matsushita Electric Ind Co Ltd 動きベクトル検出装置およびその方法
JP2000023163A (ja) * 1998-06-30 2000-01-21 Sharp Corp 動きベクトル検出装置
JP2000102015A (ja) * 1998-09-22 2000-04-07 Sharp Corp 動きベクトル検出装置及び動きベクトル検出方法

Also Published As

Publication number Publication date
JP2004023673A (ja) 2004-01-22

Similar Documents

Publication Publication Date Title
US6118488A (en) Method and apparatus for adaptive edge-based scan line interpolation using 1-D pixel array motion detection
EP0757482B1 (fr) Système de conversion de vidéo à balayage entrelacé en vidéo à balayage progressif basé sur les bords
US7671922B2 (en) Chroma upsampling method and apparatus therefor
US6266093B1 (en) Color video camera method and system
KR20010011560A (ko) 격행주사 영상의 디인터레이싱 방법
US6947094B2 (en) Image signal processing apparatus and method
JPH1188893A (ja) 画像信号処理装置
JPH08307837A (ja) テレビシステムおよび変換方法
US8355442B2 (en) Method and system for automatically turning off motion compensation when motion vectors are inaccurate
JP4510874B2 (ja) 合成映像検出装置
US8345148B2 (en) Method and system for inverse telecine and scene change detection of progressive video
JP2003339029A (ja) 動きベクトル補正回路及び方法
EP1460847B1 (fr) Appareil de traitement de signal d&#39;image et procede de traitement
US7327397B2 (en) Image processing with pixel interpolation
WO2003055211A1 (fr) Processeur de signaux d&#39;image et procede de traitement
USRE45306E1 (en) Image processing method and device thereof
WO2004002148A1 (fr) Dispositif de determination de vecteur de mouvement, procede de determination, dispositif de compensation de mouvement et procede de compensation de mouvement
US6417887B1 (en) Image display processing apparatus and method for converting an image signal from an interlaced system to a progressive system
WO2012157618A1 (fr) Dispositif de traitement de signal vidéo et dispositif d&#39;affichage
JPS62111586A (ja) 動き適応型信号処理回路
JPS6235788A (ja) ビデオ信号のフレーム周波数変換装置
JP4337302B2 (ja) 動き補正回路及び方法
JP2009124261A5 (fr)
JP2001024987A (ja) 順次走査変換回路
JP2001103433A (ja) 走査倍速信号出力方法

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): CN KR US

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT RO SE SI SK TR

DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
121 Ep: the epo has been informed by wipo that ep was designated in this application
122 Ep: pct application non-entry in european phase