JP2012032549A - Image processing device and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for improving "motion blur" while reducing an overshoot and a ghost.SOLUTION: An input image shift direction discrimination part 300 detects a positive region and a negative region. The positive region comprises pixels each having a positive gradation value when a low-frequency component image of one-previous frame is subtracted from an interested low-frequency component image. The negative region comprises pixels each having a negative gradation value when the same processing is performed. A correction processing part 440 reduces a gradation value for a part having a gradation value larger than the gradation value in a negative region of an image of one-previous sub-frame within a negative region of the selected image. Further, the correction processing part 440 reduces a gradation value for a part having a gradation value smaller than the gradation value in a positive region of an image of one-previous sub-frame within a positive region of the selected image.

Description

  The present invention relates to a technique for displaying a moving image by increasing a frame rate.

  In recent years, various devices including an image display unit such as a TV receiver and a PC monitor have been put into practical use. Various devices including a liquid crystal display device are used for these image display units.

  For example, a liquid crystal display device employs a method of dimming a backlight that is always emitting light with a liquid crystal shutter, and light is output over one frame period. Therefore, it is called a hold type display device. When a moving image is viewed in the hold-type display device (how to follow a moving part with a line of sight in moving image display), “motion blur” corresponding to the light output period is observed. For example, in moving image display at 60 Hz, in principle, a “motion blur” of at least 16.7 ms is observed.

  As a technique for reducing the “motion blur”, there is a method in which an input image with a frame rate of 60 Hz is doubled to a frame rate of 120 Hz, the spatial high-frequency component of the image is increased for one of the double-speed subframes, and the other is reduced and displayed. It has been proposed (Patent Documents 1 and 2).

JP 2006-184896 A JP 2007-304204 A

  The “motion blur” reduction technique described in the above-mentioned Patent Document 1 or Patent Document 2 is a technique that shifts the display time centroid of the spatial high-frequency component to bring the perception of the contour portion of the image closer to the impulse type. "Can be reduced. However, when the follow-up view is performed, the spatial position to be visually recognized is shifted between the image with the increased spatial high-frequency component and the image with the decreased spatial high-frequency component.

  An example in which a step change image (hereinafter simply referred to as an input image) moves from left to right at 6 pixels / frame will be described with reference to FIG. In FIG. 11, the horizontal axis indicates the pixel position, and the vertical axis indicates the pixel value (luminance value).

  An image B0 (high frequency component emphasized image) in which the spatial high frequency component of the input image shown in FIG. 11 (A) is increased and an image B1 (low frequency component image) in which the spatial high frequency component is reduced are shown in FIG. 11 (B). Shown in In the “motion blur” reduction technique described in Patent Document 1 or Patent Document 2, the frame rate of the input image is doubled, the high-frequency component emphasized image B0 is displayed in the first double-speed subframe, and low in the next double-speed subframe. The frequency component image B1 is displayed. When this display is followed, the viewpoint moves in accordance with the moving speed of the image. In the example of FIG. 11, the viewpoint moves 6 pixels to the right within one frame period.

  Assuming that the moving viewpoint is adjusted to one point, in the first double-speed subframe, the high-frequency component emphasized image B0 is visually recognized at a position shifted to the left by one pixel from the start of visual recognition. The visually recognized images (visually recognized images) are visually recognized images C0 to C2 illustrated in FIG. In the next double-speed subframe, the low-frequency component image is visually recognized at a position shifted by one pixel. The visually recognized images (visually recognized images) are visually recognized images C3 to C5 illustrated in FIG.

  Since the human eye perceives the image as the luminance obtained by time-integrating the visual images C0 to C5, the image visually recognized as a result is the visual image D shown in FIG. In addition, when the image in which the magnitude relationship of the step change gradation is inverted is used as the input image, an image obtained by vertically inverting the visual image D is visually recognized by human eyes. Further, FIG. 11 (E) shows a follow-up image E when normal display is performed without increasing or decreasing the spatial high-frequency component.

  From the comparison between the visual image D and the follow-up visual image E, the effect of reducing “motion blur” can be seen because the gradation transition width is reduced to about half by increasing or decreasing the spatial high-frequency component. However, the overshoot S caused by the high-frequency component-enhanced image is visually recognized at the gradation in front of the moving direction of the image. Further, there is a problem that a gradation step (hereinafter referred to as a ghost) G caused by a low-frequency component image is visually recognized at a gradation at the rear of the moving direction of the image.

  The present invention has been made in view of such problems, and an object of the present invention is to provide a technique for improving “motion blur” while reducing overshoot and ghost.

  In order to achieve the object of the present invention, for example, an image processing apparatus of the present invention comprises the following arrangement. That is, a generation unit that acquires a frame image input for each frame, generates a plurality of subframe images from the acquired frame image, and a low frequency component that includes a low-frequency component of the subframe image for each subframe. A generation means for generating a frequency component image and a high frequency component image in which a high frequency component of the image of the subframe is emphasized, and the high frequency component image and the low frequency component image are alternately selected in this order for each subframe. Selection means, and output means for correcting and outputting the gradation value of the image selected by the selection means, wherein the output means selects the low frequency component image of interest selected as the image of the subframe of interest by the selection means From this, a pixel whose tone value is positive when the low-frequency component image selected by the selection means is subtracted in a frame one past from the frame to which the target subframe belongs. Detecting means for detecting a region in the target low-frequency component image as a positive region and detecting a region in the target low-frequency component image composed of pixels having a negative tone value as a negative region, and the target sub In the area corresponding to the negative area in the selection image selected by the selection means in the frame to which the frame belongs, the selection means in the image selected as the image in the subframe one past from the subframe of interest. For a portion larger than the gradation value in the region corresponding to the negative region, the gradation value of the portion is reduced relative to the portion smaller than the gradation value in the region corresponding to the negative region, and the sub In the frame corresponding to the normal area in the selected image selected by the selection means in the frame to which the frame belongs, the selection means displays the image in the subframe one past from the target subframe. As for a portion smaller than the gradation value in the region corresponding to the positive region in the image selected as, the gradation value of the portion is larger than the portion in the region corresponding to the positive region. Control means for relatively reducing the output value, and means for outputting the selected image whose gradation value is controlled by the control means.

  According to the configuration of the present invention, it is possible to improve “motion blur” while reducing overshoot and ghost.

FIG. 3 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. The figure which shows the waveform of each image. The figure which shows the waveform of the image of each sub-frame, and a visual recognition image. The figure which shows the waveform of the image of each sub-frame, and a visual recognition image. 6 is a flowchart of processing performed by the image processing apparatus. The figure which shows the waveform of the image of each sub-frame, and a visual recognition image. The figure which shows the waveform of the image of each sub-frame, and a visual recognition image. The figure which shows the waveform of the image of each sub-frame, and a visual recognition image. FIG. 3 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. The figure which shows the relationship of each image. It is a figure which shows the waveform of the image of each sub-frame, and a visual recognition image.

  Preferred embodiments of the present invention will be described below with reference to the accompanying drawings. The embodiment described below shows an example when the present invention is specifically implemented, and is one of the specific examples of the configurations described in the claims.

[First Embodiment]
First, a functional configuration example of the image processing apparatus according to the present embodiment will be described with reference to the block diagram of FIG. In this embodiment, an image processing apparatus that controls gradation correction values (hereinafter simply referred to as correction values) supplied from an overdrive circuit will be described. However, the essence of this embodiment is the level of the display frame. The correction value supplied to perform the adjustment value control is controlled. Therefore, the gradation value supplier is not limited to the overdrive circuit.

  First, each part in the frame converter 100 will be described. The N-times speed processing unit 110 acquires a frame image input for each frame, and generates a plurality (N (N> 1)) of sub-frame images A from the acquired frame images. In the present embodiment, a case where N = 2 is described. In the case of N = 2, two sub-frame images are generated from one frame image, whereby a frame image input at 60 Hz can be output at 120 Hz. In addition, when N is a number larger than 2, it is easy for those skilled in the art to apply this embodiment according to the number.

  The LPF (low-pass filter) 120 cuts off (filters) the upper limit spatial frequency indicated by a predetermined constant from the sub-frame images A sequentially output from the N-times speed processing unit 110, thereby reducing the low frequency of the image A. A low-frequency component image L composed of components is generated. Note that the cutoff frequency referred to here can be considered as the number of pixels.

  The subtractor 130 subtracts the low-frequency component image L generated by the LPF 120 from the sub-frame image A (H = AL), thereby generating a high-frequency component image H that is the difference image.

  The multiplier 150 determines how many subframes the high-frequency component image H is localized and superimposes, and performs a (N−a) / a-fold operation on the high-frequency component image H. Here, a indicates the number of subframes in which the high frequency component image H is localized.

  The adder 140 adds the high-frequency component image H scaled to (N−a) / a times by the multiplier 150 and the subframe image A (SH = A + {(N−a) / a} H). = L + (N / a) H), and an image SH in which the high frequency components of the image A in the subframe are emphasized is generated. In this embodiment, since N = 2 and a = 1, SH = L + 2H.

  That is, in this embodiment, a low-frequency image L composed of low-frequency components of this sub-frame image and a high-frequency image SH that emphasizes the high-frequency components of this sub-frame image are generated from one sub-frame image. . Since this generation processing is performed for each subframe, for each subframe, a low frequency component image L composed of low frequency components of the image of this subframe and a high frequency component image SH in which the high frequency components of this subframe image are emphasized. Will be generated.

  The selector 160 switches and outputs the low-frequency component image L generated by the LPF 120 and the high-frequency component image SH generated by the adder 140 in a predetermined order in subframe units. In the present embodiment, when a subframe sequence generated from one frame is output, the high frequency component image SH is output in the first subframe, and the low frequency component image L is output in the next subframe. That is, when subframes 1 and 2 (display order) are generated from the first frame, a high-frequency component image SH is output for subframe 1 and a low-frequency component image L is output for subframe 2. When subframes 3 and 4 (display order) are generated from the next frame, a high frequency component image SH is output for subframe 3 and a low frequency component image L is output for subframe 4. As described above, the output processing is similarly performed for the subsequent frames, so that the high-frequency component image SH and the low-frequency component image L are alternately output. Similarly, even when three or more subframes are generated from one frame, the high frequency component image SH and the low frequency component image L are alternately output.

  The frame memory 200 provided outside the frame converter 100 stores a low frequency component image L selected as an output target by the selector 160. Of course, the frame memory 200 may be provided in the frame converter 100.

  In the present embodiment, one high frequency component image SH and one low frequency component image L are output from the frame converter 100 within one frame period. In the present embodiment, the low-frequency component image L output in the current frame is compared with the low-frequency component image L output in the frame one previous (past) before the current frame, so that the gradation in the current frame is The transition direction is determined. This determination process is performed by the input image transition direction determination unit 300.

  The input image transition direction discriminating unit 300 receives the low frequency component image L generated by the LPF 120 as output in the current frame F. Further, the “low frequency component image (L-1) generated by the LPF 120 to be output in frame (F-1)” stored in the frame memory 200 is input to the input image transition direction determination unit 300. The

  As the discrimination of the gradation transition direction between frame images, comparison between frame images, comparison between high-frequency component images, and difference comparison between low-frequency component images can be considered. In the comparison between the frame images, as shown in (A0) to (A2) of FIG. 2, a difference for the moving pixels of the images is obtained. However, the required transition direction determination range is a pixel range in which the high-frequency component image SH and the low-frequency component image L related to the transition of the frame image change. Therefore, a desired discrimination result cannot be obtained by comparing frame images. Further, in the comparison between the high frequency component images, as shown by (B0) to (B2), a difference between the filter range in which the high frequency component is generated and the amount of movement of the pixel is obtained. However, since the difference in this range takes both positive and negative values, the transition direction cannot be uniquely determined.

  On the other hand, in the comparison between the low frequency component images, as shown in (C0) to (C2), a difference between the filter range in which the low frequency component is generated and the amount of movement of the image is obtained. The difference value in this range is a minus value (plus at the rising transition), and the transition direction can be uniquely determined. That is, it can be determined that a pixel having a negative difference in the example of FIG. 2 is related to a falling transition between frame images. On the other hand, a pixel having a positive difference can be determined to be related to a rising transition between frame images. Here, it is important that the image used for determining the transition direction between the frame images is a low-frequency component image having the same size as the filter size of the LPF 120.

  However, in this embodiment, the input image transition direction discriminating unit 300, when the low-frequency component image (L-1) is subtracted from the low-frequency component image L, the tone value in the low-frequency component image L composed of positive pixels. When an area is detected, the rise is detected as the direction of gradation transition in this area. On the other hand, when a region in the low-frequency component image L composed of pixels having a negative tone value is detected, a fall is detected as the direction of tone transition in this region. As described above, the direction of gradation transition between the frame images can be detected from the difference between the low frequency component images.

  In the present embodiment, the transition direction determination is performed using the low-frequency component image output from the LPF 120, but a configuration other than this configuration may be employed. For example, a configuration in which determination is performed with processing equivalent to that of the LPF 120 before input to the N-times speed processing unit 110 may be employed.

  Next, each part which comprises the display gradation correction | amendment part 400 is demonstrated. The sub-frame image sent from the selector 160 (the high-frequency component image SH and the low-frequency component image L are sent alternately) includes a frame memory 410, a correction value determination unit 420, a double-speed image transition direction determination unit 500, and a correction. To the processing unit 440.

  The correction value determination unit 420 uses the image of the current subframe SF sent from the selector 160 and the image of the subframe (SF-1) read out from the frame memory 410 to calculate the level of the image of the current subframe SF. Determine the adjustment value. For example, it is assumed that the correction value determination unit 420 is an overdrive circuit or the like, or an overdrive circuit is connected to the correction value determination unit 420. In this case, the correction value determination unit 420 employs the gradation correction value determined by the overdrive circuit using the images of these two adjacent subframes. Since the technique for the overdrive circuit to determine the corresponding gradation correction value from the combination of the respective pixel values of two adjacent frame images is well known, the description thereof will be omitted.

  The double-speed image transition direction discriminating unit 500 compares the gradation value of the image of the current subframe SF with the gradation value of the image of the subframe (SF-1). Thereby, the gradation transition direction between the preceding and following subframes is determined.

  The correction value control unit 430 increases or decreases the correction value determined by the correction value determination unit 420 using the determination results of the double-speed image transition direction determination unit 500 and the input image transition direction determination unit 300, and determines the correction value. .

  The correction processing unit 440 corrects the gradation value of the image of the current subframe SF using the correction value determined by the correction value control unit 430. When the correction is completed, the correction processing unit 440 outputs the corrected subframe image.

  Next, the operation of the display gradation correction unit 400 will be described using the specific examples shown in FIGS. In FIG. 3, the content of the image moves by 6 pixels from left to right in units of frames. In FIG. 3A, the waveforms of the subframe images are listed from top to bottom in the display order. The horizontal axis indicates the pixel position, and the vertical axis indicates the gradation value.

  Next to the low-frequency component image (L-1) (second sub-frame display image of the previous input frame) output second in the frame (F-1), the high-frequency component image output first in the frame F SH (first sub-frame display image of the current input frame) is displayed. 3A, the solid line indicates the waveform of the first subframe display image of the current input frame, and the dotted line indicates the waveform of the second subframe display image of the previous input frame.

  Next to the high-frequency component image SH that is output first in the frame F, the low-frequency component image L (second sub-frame display image of the current input frame) that is output second in the frame F is displayed. 3A, the solid line indicates the waveform of the second subframe display image of the current input frame, and the dotted line indicates the waveform of the first subframe display image of the current input frame.

  FIG. 3B shows a visual image in the follow-up view when these three subframe images are sequentially displayed. It can be seen that overshoot S and ghost G are generated in this visually recognized image.

  In the case of FIG. 3, the input image transition direction discriminating unit 300 starts from a pixel whose tone value is negative when the second subframe display image of the previous input frame is subtracted from the second subframe display image of the current input frame. The negative region consisting of is detected. This negative area is an area between the pixel position 390 and the pixel position 391. If a negative region is detected, the transition direction from the previous input frame to the current input frame is a falling transition.

  In the case of FIG. 3, the double-speed image transition direction determination unit 500 compares the first subframe display image of the current input frame with the second subframe display image of the previous input frame within this negative region. As a result, in the negative area hatched portion a, the first sub-frame display image of the current input frame has a larger gradation value than the second sub-frame display image of the previous input frame, and a rising transition occurs in this portion. On the other hand, in the hatching portion b in the negative region, the gradation value of the first subframe display image of the current input frame is smaller than that of the second subframe display image of the previous input frame, and a falling transition occurs in this portion.

  Next, the double-speed image transition direction determination unit 500 compares the second subframe display image of the current input frame with the first subframe display image of the current input frame within this negative region. As a result, in the negative area hatching section c, the second sub-frame display image of the current input frame has a smaller gradation value than the first sub-frame display image of the current input frame, and this part is a falling transition. . On the other hand, in the hatched portion d of the negative region, the second subframe display image of the current input frame has a larger gradation value than the first subframe display image of the current input frame, and in this part, the transition is a rising transition.

  Here, as shown in FIG. 3B, the overshoot S in the visually recognized image is visually recognized by following the rising transition a and the falling transition c, and the ghost G is the falling transition b and the rising transition d. It is visually recognized by following vision. Therefore, in order to reduce these visual recognitions, it is achieved by controlling the transition width as (a rising transition) <(c falling transition), (b falling transition)> (d rising transition). I can do it. That is, tone correction is performed so that (falling transition)> (rising transition) in the subframe related to the falling transition of the input frame image.

  For example, the input image transition direction discriminating unit 300 and the double-speed image transition direction discriminating unit 500 make the gradation larger in the negative region in the image of the subframe SF than in the negative region in the image of the subframe (SF-1). It is assumed that portions having values (a and d in FIG. 3A) are detected. In this case, the correction value control unit 430 reduces the correction amount for the gradation value of the detected part in the image of the subframe SF.

  Then, by performing gradation correction of the corresponding portion using the reduced correction amount, an effect of attracting the overshoot S and the ghost G to the low gradation side can be obtained. And ghost G can be reduced.

  In FIG. 4, the content of the image is moved by 6 pixels from left to right in units of frames. In FIG. 4A, the waveforms of the subframe images are listed from top to bottom in the display order. The horizontal axis indicates the pixel position, and the vertical axis indicates the gradation value.

  Next to the low-frequency component image (L-1) (second sub-frame display image of the previous input frame) output second in the frame (F-1), the high-frequency component image output first in the frame F SH (first sub-frame display image of the current input frame) is displayed. 4A, the solid line indicates the waveform of the first subframe display image of the current input frame, and the dotted line indicates the waveform of the second subframe display image of the previous input frame.

  Next to the high-frequency component image SH that is output first in the frame F, the low-frequency component image L (second sub-frame display image of the current input frame) that is output second in the frame F is displayed. In the lower part of FIG. 4A, the solid line indicates the waveform of the second subframe display image of the current input frame, and the dotted line indicates the waveform of the first subframe display image of the current input frame.

  FIG. 4B shows a visually recognized image in the follow-up view when these three sub-frame images are sequentially displayed. It can be seen that overshoot S and ghost G are generated in this visually recognized image.

  In the case of FIG. 4, the input image transition direction discriminating unit 300 starts from a pixel whose tone value is positive when the second subframe display image of the previous input frame is subtracted from the second subframe display image of the current input frame. The positive region consisting of is detected. This normal area is an area between the pixel position 490 and the pixel position 491. The fact that the positive region has been detected means that the transition direction from the previous input frame to the current input frame is a rising transition.

  In the case of FIG. 4, the double-speed image transition direction determination unit 500 compares the first subframe display image of the current input frame with the second subframe display image of the previous input frame within this positive region. As a result, in the hatching portion b in the positive area, the first sub-frame display image of the current input frame has a larger gradation value than the second sub-frame display image of the previous input frame, and a rising transition occurs in this portion. On the other hand, in the hatching portion a in the positive area, the first subframe display image of the current input frame has a smaller gradation value than the second subframe display image of the previous input frame, and a falling transition occurs in this portion.

  Next, the double-speed image transition direction discriminating unit 500 compares the second subframe display image of the current input frame with the first subframe display image of the current input frame within this positive region. As a result, in the hatching portion d in the positive area, the second subframe display image of the current input frame has a smaller gradation value than the first subframe display image of the current input frame, and this portion has a falling transition. . On the other hand, in the hatching portion c in the positive area, the second sub-frame display image of the current input frame has a larger gradation value than the first sub-frame display image of the current input frame, and a rising transition occurs in this portion.

  Here, as shown in FIG. 4B, the overshoot S in the visually recognized image is visually recognized by following the rising transition c and the falling transition a, and the ghost G is the falling transition d and the rising transition b. It is visually recognized by following vision. Therefore, in order to reduce these visual recognitions, it is achieved by controlling the transition width as follows: (falling transition of a) <(rising transition of c), (rising transition of b)> (falling transition of d) I can do it. That is, gradation correction is performed so that (falling transition) <(rising transition) in the subframe related to the rising transition of the input frame image.

  For example, the input image transition direction discriminating unit 300 and the double-speed image transition direction discriminating unit 500 have a gradation smaller in the positive region in the image of the subframe SF than in the positive region in the image of the subframe (SF-1). It is assumed that portions having values (a and d in FIG. 3A) are detected. In this case, the correction value control unit 430 reduces the correction amount for the gradation value of the detected part in the image of the subframe SF.

  Then, by performing gradation correction of the corresponding portion using the reduced correction amount, an effect of attracting the overshoot S and the ghost G to the low gradation side can be obtained. And ghost G can be reduced.

  In the liquid crystal display device, since the response speed of the liquid crystal is slow, appropriate gradation display is performed by performing overdrive correction. Therefore, the gradation correction in the liquid crystal display device is realized by controlling the increase / decrease of the correction value by the overdrive circuit as described above.

  In the example of FIG. 3, the correction amount of the rising transition in the negative region is reduced, and in the example of FIG. 4, the correction amount of the falling transition in the positive region is reduced. However, the present invention is not limited to this, and the correction amount of the falling transition in the negative region may be increased in the example of FIG. 3, and the correction amount of the rising transition in the positive region may be increased in the example of FIG.

  Next, processing performed by the image processing apparatus when the above-described image processing apparatus is applied as a gradation control apparatus in a liquid crystal display device will be described with reference to FIG. 5 showing a flowchart of the processing. Note that since the flowchart of FIG. 5 shows processing for one frame, in practice, processing according to the flowchart of FIG. 5 is performed for each frame.

  In step S101, first, the N-times speed processing unit 110 generates two sub-frame images from the input frame image of the frame F. Next, for each sub-frame image, an LPF (low-pass filter) 120 generates a low-frequency component image L, and a subtractor 130, a multiplier 150, and an adder 140 generate a high-frequency component image SH.

  In step S <b> 102, the input image transition direction determination unit 300 reads “low frequency component image (L−1) output in frame (F−1)” stored in the frame memory 200.

  In step S <b> 103, the input image transition direction determination unit 300 determines, from the low frequency component image L, an area composed of pixels whose tone values are positive when the low frequency component image (L−1) is subtracted from the low frequency component image L. If detected, this region is determined to be “rise transition”. On the other hand, when a region composed of pixels with negative tone values is detected from the low-frequency component image L, it is determined that this region is a “falling transition”. If it is determined that “falling transition” is selected, the process proceeds to step S107. If it is determined that “rising transition” is selected, the process proceeds to step S104.

  In step S107, the double-speed image transition direction determination unit 500 calculates the gradation value in the negative region in the image of the current subframe SF and the gradation value in the negative region in the image of the subframe (SF-1). Compare size. When a portion having a gradation value larger than the gradation value in the negative region in the image of the subframe (SF-1) is detected in the negative region in the image of the current subframe SF, the “rise” The process proceeds to step S108. On the other hand, when a portion having a gradation value smaller than the gradation value in the negative area in the subframe (SF-1) image is detected in the negative area in the current subframe SF image, The process proceeds to step S109.

  In step S108, the correction value control unit 430 determines that “the gradation value in the negative region in the image of the current subframe SF is larger than the gradation value in the negative region in the image of the subframe (SF-1). The correction value determined by the correction value determination unit 420 for “part” is reduced. Of course, the correction value may be zero.

  In step S109, the correction value control unit 430 “has a gradation value smaller than the gradation value in the negative region in the image of the subframe (SF-1) in the negative region in the image of the current subframe SF. The correction value determined by the correction value determination unit 420 for “part” is increased.

  On the other hand, in step S104, the double-speed image transition direction determination unit 500 determines the gradation value in the positive region in the image of the current subframe SF and the gradation value in the positive region in the image of the subframe (SF-1). Compare the size of. When a portion having a gradation value larger than the gradation value in the positive region in the image of the subframe (SF-1) is detected in the positive region in the image of the current subframe SF, “rise” The process proceeds to step S105. On the other hand, when a portion having a gradation value smaller than the gradation value in the positive region in the subframe (SF-1) image is detected in the positive region in the current subframe SF image, The process proceeds to step S106.

  In step S106, the correction value control unit 430 “has a gradation value smaller than the gradation value in the positive region in the image of the subframe (SF-1) in the positive region in the image of the current subframe SF. The correction value determined by the correction value determination unit 420 for “part” is reduced. Of course, the correction value may be zero.

  In step S <b> 105, the correction value control unit 430 determines that “in the positive region in the image of the current subframe SF, the gradation value is larger than the gradation value in the positive region in the image of the subframe (SF-1). The correction value determined by the correction value determination unit 420 for “part” is increased.

  Thereafter, the correction processing unit 440 performs tone correction using a correction value determined for the positive region for the positive region in the image of the current subframe SF, and for the negative region, The gradation correction is performed using the correction value determined for the negative region. When the correction is completed, the correction processing unit 440 outputs the corrected subframe image.

  As described above, according to the present embodiment, the gradation transition direction in the current frame can be determined from the difference between the low frequency component image generated for the previous input frame and the low frequency component image generated for the current frame. I can do it. Then, gradation correction of gradation transition in the subframe is performed according to the determined gradation transition direction.

  That is, the main points of this embodiment are as follows. A frame image input for each frame is acquired, and images of a plurality of subframes are generated from the acquired frame image. Then, for each sub-frame, a low-frequency component image composed of low-frequency components of the sub-frame image and a high-frequency component image in which the high-frequency component of the sub-frame image is emphasized are generated. Then, the high-frequency component image and the low-frequency component image are alternately selected for each subframe in the selection order of the high-frequency component image and the low-frequency component image, and the gradation value of the selected image is corrected and output. In this output, the following operation is performed.

  First, gradation is obtained when the low-frequency component image selected in the frame (F-1) one frame before the frame F to which the target subframe SF belongs is subtracted from the low-frequency component image of the target subframe SF. A region in the low-frequency component image of interest consisting of pixels having a positive value is detected. Similarly, a region in the low frequency component image of interest composed of pixels with negative tone values is detected. Then, in the region corresponding to the negative region in the selected image selected in the frame F, from the gradation value in the region corresponding to the negative region in the image selected as the image of the subframe (SF-1). For the large part, the following processing is performed. That is, the gradation value of this portion is reduced with respect to a portion smaller than the gradation value in the region corresponding to the negative region. On the other hand, in the region corresponding to the positive region in the selected image selected in the frame F, from the gradation value in the region corresponding to the positive region in the image selected as the image of the subframe (SF-1). For the smaller part, the following processing is performed. That is, the gradation value of this part is relatively reduced with respect to the part larger than the gradation value in the area corresponding to the positive area. Then, the selected image whose gradation value is controlled is output. Thereby, it is possible to improve “motion blur” while reducing overshoot and ghost.

[Second Embodiment]
In the first embodiment, a high frequency component image is first output and then a low frequency component image is output within one frame. In this embodiment, the low frequency component image is output first, and then the high frequency component image is output. Only the parts different from the first embodiment will be described below.

  In FIG. 6, the content of the image is moved by 6 pixels from left to right in units of frames. In FIG. 6A, the waveforms of the subframe images are listed from top to bottom in the display order. The horizontal axis indicates the pixel position, and the vertical axis indicates the gradation value.

  Next to the high-frequency component image (SH-1) (second sub-frame display image of the previous input frame) output second in the frame (F-1), the low-frequency component image output first in the frame F L (first sub-frame display image of the current input frame) is displayed. 6A, the solid line indicates the waveform of the first subframe display image of the current input frame, and the dotted line indicates the waveform of the second subframe display image of the previous input frame.

  Next to the low-frequency component image L output first in the frame F, the high-frequency component image SH output second in the frame F (second sub-frame display image of the current input frame) is displayed. 6A, the solid line indicates the waveform of the second subframe display image of the current input frame, and the dotted line indicates the waveform of the first subframe display image of the current input frame.

  FIG. 6B shows a visual image in the follow-up view when these three sub-frame images are sequentially displayed. It can be seen that the overshoot S and the ghost G in the visually recognized image are different from those in the first embodiment.

  In the case of FIG. 6, the input image transition direction determination unit 300 has a negative gradation value when the first subframe display image (not shown) of the previous input frame is subtracted from the first subframe display image of the current input frame. The negative region consisting of the pixels to be detected is detected. If a negative region is detected, the transition direction from the previous input frame to the current input frame is a falling transition.

  In the case of FIG. 6, the double-speed image transition direction discriminating unit 500 compares the first subframe display image of the current input frame with the second subframe display image of the previous input frame within this negative region. As a result, in the hatched portion a in the negative region, the gradation value of the first subframe display image of the current input frame is smaller than that of the second subframe display image of the previous input frame. is there. On the other hand, in the hatched portion b in the negative region, the first sub-frame display image of the current input frame has a larger gradation value than the second sub-frame display image of the previous input frame, and a rising transition occurs in this portion.

  Next, the double-speed image transition direction determination unit 500 compares the second subframe display image of the current input frame with the first subframe display image of the current input frame within this negative region. As a result, in the hatched portion c in the negative area, the second subframe display image of the current input frame has a larger gradation value than the first subframe display image of the current input frame, and in this part, the transition is a rising transition. . On the other hand, in the hatched portion d in the negative region, the second sub-frame display image of the current input frame has a smaller gradation value than the first sub-frame display image of the current input frame, and a falling transition occurs in this portion. .

  Here, as shown in FIG. 6B, the ghost G in the visually recognized image is visually recognized by following the rising transition c and the falling transition a, and the overshoot S is the falling transition d and the rising transition b. It is visually recognized by following vision. Therefore, in order to reduce these visual recognitions, it is achieved by controlling the transition width as follows: (falling transition of a) <(rising transition of c), (rising transition of b)> (falling transition of d) I can do it. That is, tone correction is performed so that (falling transition) <(rising transition) in the subframe related to the falling transition of the input frame image.

  For example, the input image transition direction discriminating unit 300 and the double-speed image transition direction discriminating unit 500 make the gradation smaller in the negative region in the image of the subframe SF than in the negative region in the image of the subframe (SF-1). It is assumed that portions having values (a and d in FIG. 6A) are detected. In this case, the correction value control unit 430 reduces the correction amount for the gradation value of the detected part in the image of the subframe SF.

  Then, by performing gradation correction of the corresponding portion using the reduced correction amount, an effect of attracting the overshoot S and the ghost G to the high gradation side can be obtained. The ghost G can be reduced.

  In FIG. 7, the content of the image is moved by 6 pixels from left to right in units of frames. In FIG. 7A, the waveforms of the subframe images are listed from top to bottom in the display order. The horizontal axis indicates the pixel position, and the vertical axis indicates the gradation value.

  Next to the high-frequency component image (SH-1) (second sub-frame display image of the previous input frame) output second in the frame (F-1), the low-frequency component image output first in the frame F L (first sub-frame display image of the current input frame) is displayed. 7A, the solid line indicates the waveform of the first subframe display image of the current input frame, and the dotted line indicates the waveform of the second subframe display image of the previous input frame.

  Next to the low-frequency component image L output first in the frame F, the high-frequency component image SH output second in the frame F (second sub-frame display image of the current input frame) is displayed. 7A, the solid line indicates the waveform of the second subframe display image of the current input frame, and the dotted line indicates the waveform of the first subframe display image of the current input frame.

  FIG. 7B shows a visual image in the follow-up view when these three sub-frame images are sequentially displayed. It can be seen that the overshoot S and the ghost G in the visually recognized image are different from those in the first embodiment.

  In the case of FIG. 7, the input image transition direction discriminating unit 300 corrects the gradation value when the first subframe display image (not shown) of the previous input frame is subtracted from the first subframe display image of the current input frame. The positive region consisting of the pixels to be detected is detected. The fact that the positive region has been detected means that the transition direction from the previous input frame to the current input frame is a rising transition.

  In the case of FIG. 7, the double-speed image transition direction discriminating unit 500 compares the first subframe display image of the current input frame with the second subframe display image of the previous input frame within this positive region. As a result, in the hatched portion a in the positive region, the first sub-frame display image of the current input frame has a larger gradation value than the second sub-frame display image of the previous input frame, and this part is a rising transition. . On the other hand, in the hatching portion b in the positive region, the first subframe display image of the current input frame has a smaller gradation value than the second subframe display image of the previous input frame, and this part is a falling transition. .

  Next, the double-speed image transition direction discriminating unit 500 compares the second subframe display image of the current input frame with the first subframe display image of the current input frame within this positive region. As a result, in the hatched portion c in the positive region, the second subframe display image of the current input frame has a smaller gradation value than the first subframe display image of the current input frame, and in this part, the transition is caused by a falling transition. is there. On the other hand, in the hatched portion d in the positive area, the second sub-frame display image of the current input frame has a larger gradation value than the first sub-frame display image of the current input frame, and a rising transition occurs in this portion.

  Here, as shown in FIG. 7B, the ghost G in the visual image is visually recognized by following the rising transition a and the falling transition c, and the overshoot S is the falling transition b and the rising transition d. It is visually recognized by following vision. Therefore, in order to reduce these visual recognitions, it is achieved by controlling the transition width as (a rising transition) <(c falling transition), (b falling transition)> (d rising transition). I can do it. That is, gradation correction is performed so that (falling transition)> (rising transition) in the subframe related to the rising transition of the input frame image.

  For example, the input image transition direction discriminating unit 300 and the double-speed image transition direction discriminating unit 500 have a gradation larger in the positive region in the image of the subframe SF than in the positive region in the image of the subframe (SF-1). It is assumed that portions having values (a and d in FIG. 6A) are detected. In this case, the correction value control unit 430 reduces the correction amount for the gradation value of the detected part in the image of the subframe SF.

  Then, by performing gradation correction of the corresponding portion using the reduced correction amount, an effect of attracting the overshoot S and the ghost G to the low gradation side can be obtained. And ghost G can be reduced.

  As described above, according to the present embodiment, even when the output from the frame converter is in the order of the low-frequency component image and the high-frequency component image, it is possible to improve “motion blur” while reducing overshoot and ghost. .

[Third Embodiment]
In the first and second embodiments, the case where the frame converter 100 divides a frame input at 60 Hz into two sub-subframes (N = 2), converts the frame into a 120 Hz image, and outputs the image. In this embodiment, N = 3, and three subframes generated from one frame are sequentially output. In this case, the low frequency component image generated from this subframe is output in the first subframe, the high frequency component image generated from this subframe is output in the next subframe, and this subframe is output in the last subframe. The low frequency component image generated from is output. Only the parts different from the first embodiment will be described below.

  In FIG. 8, the content of the image is moved by 6 pixels from left to right in units of frames. In FIG. 8A, the waveforms of the subframe images are listed from top to bottom in the display order. The horizontal axis indicates the pixel position, and the vertical axis indicates the gradation value.

  Next to the low-frequency component image output third in the frame (F-1) (third sub-frame display image of the previous input frame), the low-frequency component image output first in the frame F (the current input frame). The first sub-frame display image) is displayed. In the second row from the top in FIG. 8A, the solid line indicates the waveform of the first subframe display image of the current input frame, and the dotted line indicates the waveform of the third subframe display image of the previous input frame.

  Secondly, in the frame F, a high-frequency component image (second sub-frame display image of the current input frame) is displayed. In the third row from the top in FIG. 8A, the solid line indicates the waveform of the second subframe display image of the current input frame, and the dotted line indicates the waveform of the first subframe display image of the current input frame.

  In the third frame F, a low-frequency component image (third sub-frame display image of the current input frame) is displayed. In the fourth row from the top in FIG. 8A, the solid line indicates the waveform of the third subframe display image of the current input frame, and the dotted line indicates the waveform of the second subframe display image of the current input frame.

  FIG. 8B shows a visual image in the follow-up view when these four subframe images are sequentially displayed. As shown in FIG. 8B, in the visually recognized image, the overshoot in the follow-up view is not visible, and the ghosts G0 and G1 are visible.

  In FIG. 8A, the transition direction from the third subframe display image of the previous input frame to the first subframe display image of the current input frame is a falling transition in the hatching portion a. Also, from the first subframe display image of the current input frame to the second subframe display image of the current input frame, the transition direction is a rising transition in the hatching portion b, and the transition direction is a falling transition in the hatching portion c. It is. Also, from the second subframe display image of the current input frame to the third subframe display image of the current input frame, the transition direction is a rising transition in the hatching portion e, and the transition direction is a falling transition in the hatching portion d. It is.

  Here, the ghost G0 in the visually recognized image is visually recognized by the tracking of the falling transition a and the rising transition b, and the ghost G1 is visually recognized by the tracking of the falling transition c and the rising transition e. Therefore, in order to reduce these visual recognitions, it is achieved by controlling the transition width as (falling transition of a) <(rising transition of b), (falling transition of c)> (rising transition of e). I can do it.

  By the way, in this embodiment, regarding the falling transition of the frame image, a case where the rising transition of the subframe is strengthened and a case where the falling transition is strengthened are mixed. In this case, the type of output image from the frame converter 100 is used for determination. That is, when a low-frequency component image is displayed in the previous subframe, gradation correction is performed so that (falling transition) <(rising transition). Further, when a high frequency component image is displayed in the previous subframe, gradation correction is performed so that (falling transition)> (rising transition).

  By this correction, it is possible to obtain an effect of raising the ghost G0 to the high gradation side and lowering the ghost G1 to the low gradation side, and as a result, the ghosts G0 and G1 can be reduced. Note that the type of the output image from the frame converter 100 may be determined by adding information as a flag to the output signal from the frame converter 100.

  As described above, according to the present embodiment, the gradation correction method in the gradation transition of the subframe can be switched according to the type of the output image from the frame converter and the gradation transition direction of the frame image. it can. As a result, even when triple-speed display is used, it is possible to improve “motion blur” while reducing ghost.

[Fourth Embodiment]
A functional configuration example of the image processing apparatus according to the present embodiment will be described with reference to the block diagram of FIG. The configuration shown in FIG. 9 is the same as the configuration shown in FIG. 1 except that the frame memory 410 is omitted from the configuration shown in FIG. 1 and a selector 210 is added instead. The selector 210 alternately switches and outputs the high frequency component image SH from the adder 140 and the low frequency component image L from the frame memory 200.

  An image output from the frame converter 100, each image generated inside the frame converter 100, and an image read from the frame memory 200 will be described with reference to FIG.

  For example, in the frame F1, a high frequency component image H1 and a low frequency component image L1 are generated in the frame converter 100. Then, the selector 160 switches and outputs the high frequency component image H1 in the first subframe and the low frequency component image L1 in the next subframe. In the frame F1, the low frequency component image L0 generated in the previous frame F0 is read from the frame memory 200. The input image transition direction determination unit 300 determines the transition direction from the read low frequency component image L0 and the current low frequency component image L1.

  The selector 210 selects the low frequency component image L0 read from the frame memory 200 in the first subframe in the frame F1, and selects the high frequency component image H1 generated by the frame converter 100 in the next subframe. Output. By this control, the correction value determination unit 420 and the double speed image transition direction determination unit 500 determine the transition direction for the gradation transition from the low frequency component image L0 to the high frequency component image H1 in the first subframe in the frame F1. Further, the correction value can be determined. In the next subframe, it is possible to determine the transition direction and determine the correction value for the gradation transition from the high frequency component image H1 to the low frequency component image L1. In the present embodiment, the output from the frame converter 100 is in the order of a high-frequency component image in the first subframe and a low-frequency component image in the next subframe. As described above, according to the present embodiment, it is possible to improve “motion blur” while reducing overshoot and ghost with a configuration using a single frame memory.

[Fifth Embodiment]
Each image processing described in the first to fourth embodiments is realized by the operation of hardware having the configuration shown in FIGS. 1 and 9 in each embodiment. However, a computer program for causing the CPU to execute each image processing is installed in a memory in a computer having the CPU, and the CPU executes the computer program, thereby causing the computer to realize the image processing. May be. A general personal computer or the like can be applied to such a computer.

(Other examples)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

Claims (7)

Generating means for acquiring a frame image input for each frame and generating images of a plurality of subframes from the acquired frame image;
Generating means for generating, for each subframe, a low frequency component image composed of low frequency components of the image of the subframe, and a high frequency component image in which the high frequency component of the image of the subframe is emphasized;
Selection means for alternately selecting the high-frequency component image and the low-frequency component image in this order for each subframe;
Output means for correcting and outputting the gradation value of the image selected by the selection means,
The output means includes
When the low-frequency component image selected by the selection unit in the frame one past from the frame to which the target sub-frame belongs is subtracted from the target low-frequency component image selected by the selection unit as the image of the target sub-frame. An area in the target low-frequency component image composed of pixels with a positive gradation value is detected as a positive area, and an area in the target low-frequency component image composed of pixels with a negative gradation value is detected as a negative area. Detecting means for
An image selected by the selection unit as an image in a subframe one previous to the target subframe in a region corresponding to the negative region in the selection image selected by the selection unit in a frame to which the target subframe belongs. For a portion larger than the gradation value in the region corresponding to the negative region, the gradation value of the portion is reduced with respect to a portion smaller than the gradation value in the region corresponding to the negative region,
An image selected by the selection unit as an image in a subframe one previous to the target subframe in an area corresponding to the positive region in the selected image selected by the selection unit in a frame to which the target subframe belongs. For a portion smaller than the gradation value in the region corresponding to the positive region, the gradation value of the portion is set relatively to the portion larger than the gradation value in the region corresponding to the positive region. Control means to reduce,
An image processing apparatus comprising: means for outputting the selected image whose gradation value is controlled by the control means. Generating means for acquiring a frame image input for each frame and generating images of a plurality of subframes from the acquired frame image;
Generating means for generating, for each subframe, a low frequency component image composed of low frequency components of the image of the subframe, and a high frequency component image in which the high frequency component of the image of the subframe is emphasized;
Selection means for alternately selecting the low-frequency component image and the high-frequency component image in this order for each subframe;
Output means for correcting and outputting the gradation value of the image selected by the selection means,
The output means includes
When the low-frequency component image selected by the selection unit in the frame one past from the frame to which the target sub-frame belongs is subtracted from the target low-frequency component image selected by the selection unit as the image of the target sub-frame. An area in the target low-frequency component image composed of pixels with a positive gradation value is detected as a positive area, and an area in the target low-frequency component image composed of pixels with a negative gradation value is detected as a negative area. Detecting means for
An image selected by the selection unit as an image in a subframe one previous to the target subframe in a region corresponding to the negative region in the selection image selected by the selection unit in a frame to which the target subframe belongs. For a portion smaller than the gradation value in the region corresponding to the negative region, the gradation value of the portion is reduced relative to the portion smaller than the gradation value in the region corresponding to the negative region,
An image selected by the selection unit as an image in a subframe one previous to the target subframe in an area corresponding to the positive region in the selected image selected by the selection unit in a frame to which the target subframe belongs. For a portion larger than the gradation value in the region corresponding to the positive region, the gradation value of the portion is set relatively to the portion larger than the gradation value in the region corresponding to the positive region. Control means to reduce,
An image processing apparatus comprising: means for outputting the selected image whose gradation value is controlled by the control means.   The image processing apparatus according to claim 1, wherein the control unit controls a gradation value supplied from an overdrive circuit. Generating a frame image input for each frame, and generating a plurality of sub-frame images from the acquired frame image;
A generating step for generating, for each subframe, a low-frequency component image composed of low-frequency components of the image of the subframe and a high-frequency component image in which the high-frequency component of the image of the subframe is emphasized;
A selection step of alternately selecting the high-frequency component image and the low-frequency component image in this order for each subframe;
An output step of correcting and outputting the gradation value of the image selected in the selection step,
The output step includes
When the low-frequency component image selected in the selection step is subtracted from the target low-frequency component image selected as the image of the target sub-frame in the selection step within a frame one past from the frame to which the target sub-frame belongs. An area in the target low-frequency component image composed of pixels with a positive gradation value is detected as a positive area, and an area in the target low-frequency component image composed of pixels with a negative gradation value is detected as a negative area. Detecting step to
An image selected as an image in a subframe one previous to the target subframe in the selection step in a region corresponding to the negative region in the selection image selected in the selection step in the frame to which the target subframe belongs. For a portion larger than the gradation value in the region corresponding to the negative region, the gradation value of the portion is reduced with respect to a portion smaller than the gradation value in the region corresponding to the negative region,
An image selected as an image in a subframe one previous to the target subframe in the selection step in a region corresponding to the positive region in the selected image selected in the selection step in the frame to which the target subframe belongs. For a portion smaller than the gradation value in the region corresponding to the positive region, the gradation value of the portion is set relatively to the portion larger than the gradation value in the region corresponding to the positive region. Control process to reduce,
And a step of outputting the selected image whose gradation value is controlled in the control step. Generating a frame image input for each frame, and generating a plurality of sub-frame images from the acquired frame image;
A generating step for generating, for each subframe, a low-frequency component image composed of low-frequency components of the image of the subframe and a high-frequency component image in which the high-frequency component of the image of the subframe is emphasized;
A selection step of alternately selecting the low-frequency component image and the high-frequency component image in this order for each subframe;
An output step of correcting and outputting the gradation value of the image selected in the selection step,
The output step includes
When the low-frequency component image selected in the selection step is subtracted from the target low-frequency component image selected as the image of the target sub-frame in the selection step within a frame one past from the frame to which the target sub-frame belongs. An area in the target low-frequency component image composed of pixels with a positive gradation value is detected as a positive area, and an area in the target low-frequency component image composed of pixels with a negative gradation value is detected as a negative area. Detecting step to
An image selected as an image in a subframe one previous to the target subframe in the selection step in a region corresponding to the negative region in the selection image selected in the selection step in the frame to which the target subframe belongs. For a portion smaller than the gradation value in the region corresponding to the negative region, the gradation value of the portion is reduced relative to the portion smaller than the gradation value in the region corresponding to the negative region,
An image selected as an image in a subframe one previous to the target subframe in the selection step in a region corresponding to the positive region in the selected image selected in the selection step in the frame to which the target subframe belongs. For a portion larger than the gradation value in the region corresponding to the positive region, the gradation value of the portion is set relatively to the portion larger than the gradation value in the region corresponding to the positive region. Control process to reduce,
And a step of outputting the selected image whose gradation value is controlled in the control step.   The computer program for functioning a computer as each means which the image processing apparatus of any one of Claims 1 thru | or 3 has.   A computer-readable storage medium storing the computer program according to claim 6.

Images