JP2009258343A - Image display device and image display method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve both video pseudo contour and video blur in a sub-field emission type image display device by suppressing false color when an object non-uniformly moves. <P>SOLUTION: A motion vector detection part 11 detects a motion vector of a pixel between frames, and a histogram count part 12 determines a distribution of motion vectors for each frame. A pixel position switching part 15 calculates and outputs a pixel position vector Xi showing an acquisition destination of light emission data by use of a motion vector having a pixel to be reconfigured as an end point and an arithmetic expression. At that time, when the distribution of motion vectors is not uniformed within a frame, the pixel position vector Xi is corrected in reference to luminance information so as to be close to the pixel to be reconfigured until a luminance difference between the pixel to be reconfigured and the pixel of the acquisition destination is a threshold or less, and outputs a new pixel position vector Yi. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image display apparatus and an image display method for performing gradation display by time-dividing one frame (or one field) in image display.

  One frame is divided into a plurality of screens with different luminance weights (hereinafter referred to as subfields (SF)) in the time direction, and one frame image is controlled by controlling light emission and non-light emission in each subfield. In a display device that displays a moving image, there is a problem in that a gradation disturbance called moving image pseudo contour or moving image blur occurs, and the display quality is deteriorated. This is known to occur because the human eye follows a moving object.

  In order to prevent the occurrence of the moving image pseudo contour, Patent Document 1 discloses that a motion vector is detected from display data between frames or fields, and a light emission position of each subfield of the display data is calculated from the motion vector. A method for correcting the pixel position of each subfield on the line-of-sight path has been proposed.

  Further, Patent Document 2 discloses a method of calculating a drag coordinate indicating where each subfield of the current pixel comes from the motion vector and the emission barycentric position of the subfield, and re-encoding the subfield. Proposed.

  Patent Document 3 relates to correction of a motion vector used for motion blur suppression image processing in a hold-type display device such as a liquid crystal display device, and the frequency of motion vector candidates in a pixel of interest and its surrounding pixels. Describes that the motion vector candidate having the highest value is determined as the motion vector of the pixel of interest, and that the motion vector is corrected based on the degree of change in luminance around the pixel of interest.

JP-A-8-2111848 Japanese Patent Laid-Open No. 2002-123211 JP 2005-160015 A

  According to the techniques of Patent Documents 1 and 2 described above, the moving image pseudo contour can be effectively removed when the motion vector is known, such as when the entire screen moves in the same direction. Moreover, according to the technique of the said patent document 3, it is described that the motion blur resulting from the human follow-up vision in a hold type | mold display apparatus can be suppressed by detecting the motion vector in harmony with the circumference | surroundings. . However, this technique may correct the motion vector to be significantly larger than the actual magnitude of the motion, and the present inventors are also able to correct the entire screen if the movement of the entire screen is not uniform. Similar to the case of moving in the same direction, a problem has been found that when a correction is made using a motion vector, a new pseudo contour is generated.

  That is, since the motion vector is obtained for each object in the screen and the light emission position of the subfield is corrected, when motion vectors of various directions and sizes are detected in each object, the motion vector is used. When the light emission position of the subfield is corrected, the color change of the pixels in the peripheral area of the moving object is increased, and a false color is generated.

  In addition, when the light emission position of the subfield is corrected so as to suppress a large color change of pixels in the peripheral area of the moving object, there is a problem that the effect of improving the motion blur is small.

  First, FIG. 12 is a diagram illustrating a gray scale expression method of a display device that expresses gray scales using subfields. One frame is composed of N subfields, and weighting such as 2 to the Nth power is performed in each subfield. In this example, weighting is performed with 2 to the 0th power, 2 to the 1st power,. Called SF1, SF2,..., SFN from the start side of one frame period. Here, N = 8 is an example. In the display device, gradation within one frame is expressed by selecting a plurality of light emission and non-light emission in this subfield. The luminance perceived by the human retina is the sum of the luminance values of a plurality of subfields that emit light.

  Here, since the light emission of the subfields is temporally different, when the human eye follows a moving object in the moving image and the position of the light emission subfield of an adjacent pixel in one frame changes greatly, a moving image pseudo contour Occurs.

  FIG. 13 shows an example of a mechanism for generating a moving image pseudo contour. The vertical direction represents time, the horizontal direction represents the pixel position, the number N of subfields is 8, and the horizontal direction represents a case where a series of pixels each having one pixel in the left direction and having a high luminance are displayed.

  FIG. 13A shows a case where the series of pixel displays are moved to the right by two pixels in the second frame period than in the first frame period. Here, the pixels having the luminances 127, 128, and 129 shown in the drawing are originally visible to the human eyes as 127, 128, and 129, respectively, in a still image state.

  However, in the case of a moving image, the line of sight follows the movement of the image as shown by the arrows in the figure. Thereby, the light emission period of the subfield recognized by human eyes is different from that in the case of a still image. In the example of (a), pixels having luminance of 127, 128, and 129 are recognized by human eyes for still images, and pixels having luminance of 127, 0, and 129 are displayed for moving images. In this way, the human eye recognizes a pixel with 0 brightness that should not be displayed.

  In addition, as shown in FIG. 13B, when the series of pixel displays are moved to the left by two pixels in the second field period than in the first field period, the luminance is increased in the case of a still image. The 126, 127, and 128 pixels are recognized by the human eye as pixels having luminance of 126, 255, and 128 for moving image display. In this way, the human eye recognizes a pixel having a luminance of 255 that should not be displayed. This is the mechanism of moving image pseudo contour generation.

  In view of the above problems, an object of the present invention is to suppress false color when the motion of an object in a screen is not uniform in a subfield light emitting type image display device, and improve both moving image pseudo contour and moving image blur. That is.

  The present invention divides one frame of an input image into a plurality of subfield periods and reconstructs light emission data in each period of the plurality of subfield periods according to the motion vectors of corresponding pixels between the frames. An apparatus comprising: a subfield conversion unit that converts an input image into light emission data of a plurality of subfields; a motion vector detection unit that detects a motion vector of a pixel between frames of the input image; and a distribution of detected motion vectors. A histogram count unit obtained for each frame, a luminance information calculation unit that calculates luminance information of each pixel from the input image, and a motion vector having the end point of the reconstruction target pixel in the reconstruction target frame among the detected motion vectors Acquisition of data for selecting and reconstructing light emission data from the selected motion vector using a predetermined arithmetic expression A pixel position switching unit that calculates and outputs a pixel position vector indicating the output, and the pixel field output from the pixel position switching unit for the emission data of the subfield of each pixel in the reconstruction target frame output from the subfield conversion unit An image using the subfield reconstruction unit reconstructed using the light emission data of the corresponding subfield of the pixel in the reconstruction target frame indicated by the position vector, and the subfield light emission data output from the subfield reconstruction unit The pixel position switching unit corrects the calculated pixel position vector for each frame in accordance with the motion vector distribution obtained by the histogram count unit and the luminance information calculated by the luminance information calculation unit. And output.

  Here, the pixel position switching unit outputs the calculated pixel position vector Xi when the motion vector distribution obtained by the histogram count unit is uniform within the frame, and the motion vector distribution is not uniform within the frame. Refers to the luminance information calculated by the luminance information calculation unit, and calculates the calculated pixel position vector Xi until a pixel whose luminance difference between the pixel indicated by the pixel position vector Xi and the pixel to be reconstructed is equal to or smaller than a threshold value appears. The pixel position vector Yi is corrected so as to be close to the reconfiguration target pixel and is output.

  And a motion vector correction unit that corrects the motion vector V detected by the motion vector detection unit to another motion vector based on the motion vector distribution obtained by the histogram count unit. When the motion vector distribution obtained by the unit is uniform within the frame, each motion vector V within the frame detected by the motion vector detection unit is replaced with a non-zero motion vector having the maximum count value Nn within the frame. If the motion vector distribution is not uniform within the frame, the detected motion vector V is used as it is.

  The present invention divides one frame of an input image into a plurality of subfield periods and reconstructs light emission data in each period of the plurality of subfield periods according to the motion vectors of corresponding pixels between the frames. A method comprising: converting an input image into light emission data of a plurality of subfields; detecting a motion vector of a pixel between frames in the input image; obtaining a distribution of the detected motion vector for each frame; Calculating the luminance information of each pixel from the input image; and selecting a motion vector having the end point of the reconstruction target pixel in the reconstruction target frame from the detected motion vectors, and performing a predetermined calculation from the selected motion vector The pixel position vector indicating the data acquisition source for reconstructing the emission data is calculated using the equation, and the motion vector is calculated. A step of correcting and outputting the calculated pixel position vector for each frame in accordance with the distribution and luminance information of the image, and a reconstruction in which the pixel position vector indicates the emission data of the subfield of each pixel in the reconstruction target frame Reconstructing using the light emission data of the corresponding subfield of the pixel in the target frame, and displaying the image using the light emission data of the subfield.

  According to the present invention, pseudo contour and moving image blur are reduced while suppressing false colors when the motion of an object in the screen is not uniform. Therefore, it is possible to provide a high-quality image without image quality deterioration.

  Embodiments of the present invention will be described below with reference to the drawings. In each drawing, the component to which the same code | symbol is attached | subjected shall have the same function. In the following description, the description “subfield” also includes the meaning of “subfield period”. In addition, the description “light emission of the subfield” includes the meaning of “light emission of the pixel in the subfield period”. In addition, in the following description or drawings, when a scalar quantity is simply described as the value of a motion vector, it is assumed that the motion quantity in the horizontal direction of the two-dimensional vector is exemplified. For example, the notation “6” simply indicates that the motion vector is (x, y) = (+ 6, 0) when the horizontal direction of the display screen is x and the vertical direction is y.

  FIG. 1 is a block diagram showing an image display apparatus according to a first embodiment of the present invention. The configuration of the image display device 1 includes an input unit 10, a motion vector detection unit 11, a histogram count unit 12, a subfield conversion unit 13, a luminance information calculation unit 14, a pixel position switching unit 15, a subfield reconstruction unit 16, and an image display. A unit 17 and a control unit 18 are provided.

  The operation of each part will be described. Moving image data is input to the input unit 10. For example, the input unit includes a TV broadcast tuner, an image input terminal, a network connection terminal, and the like. The input unit 10 performs conventional conversion processing or the like on the input moving image data, and outputs display data after the conversion processing to the motion vector detection unit 11.

  The motion vector detection unit 11 compares the display data of the target frame with the display data of the frame temporally prior to the target frame, thereby obtaining a motion vector V (reference numeral 101) that ends each pixel of the target frame. To detect. From this motion vector, information on the speed and direction of movement of the object can be obtained. Here, when the horizontal component of a motion vector of a pixel is Vx and the vertical component is Vy, the motion vector of the pixel is represented as V = (Vx, Vy). Regarding the motion detection technique and the motion estimation technique for detecting the motion vector, since a well-known technique used in the MPEG encoding process or the like can be applied, description thereof is omitted here.

  The histogram count unit 12 counts the frequency of appearance of the motion vector V (101) detected by the motion vector detection unit 11 for each of the horizontal component Vx and the vertical component Vy, and outputs a motion vector distribution (histogram information) 102. To do. The histogram shows how many motion vectors having each component exist in a certain area in a distribution diagram. When the motion vector distribution is concentrated on a specific motion vector, it is expressed that “bias” occurs in the specific motion vector. That is, “bias” in a specific motion vector means that the motion of the image in the region is uniform.

  The subfield conversion unit 13 converts the input image into light emission data of a plurality of subfields. The luminance information calculation unit 14 calculates luminance information 103 from the video data input to the input unit 10.

  The pixel position switching unit 15 calculates a pixel position vector indicating the pixels of the subfield before reconstruction in order to rearrange the subfields of the target pixel. At this time, the motion vector 101 detected by the motion vector detection unit 11, the motion vector distribution (histogram information) 102 obtained by the histogram count unit 12, and the luminance information 103 calculated by the luminance information calculation unit 14 are used. The threshold value (threshold value) S (reference numeral 104) and the threshold value (threshold value) T (reference numeral 105) are used to determine whether or not the motion vector distribution is uniform.

  First, the pixel position switching unit 15 receives the motion vector 101 detected by the motion vector detection unit 11 and calculates a pixel position vector Xi using an arithmetic expression. Next, the luminance information 103 calculated by the luminance information calculation unit 14 is used to determine the luminance difference between the pixel indicated by the pixel position vector Xi and the pixel to be reconfigured. A new pixel position vector Yi that is equal to or less than the threshold is obtained. Further, the motion vector distribution 102 obtained by the histogram count unit 12 is determined for each frame, and when the distribution is uniform, the pixel position vector Xi is switched, and when the distribution is not uniform, the pixel position vector Yi is switched and output. .

  The subfield reconstruction unit 16 obtains light emission data of the subfield of the pixel indicated by the pixel position vector output by switching for each frame by the pixel position switching unit 15 among the subfield data output by the subfield conversion unit 13. . The acquired light emission data is arranged in the subfield to be reconfigured and reconfiguration is performed. By repeating this for each subfield, the subfield is reconstructed for each pixel, and the subfield data output from the subfield conversion unit 13 is reconstructed.

  The image display unit 17 includes a plurality of pixels that perform light emission operations such as lighting and extinguishing, and displays an image by controlling lighting or extinguishing of each pixel based on the subfield data obtained by the subfield reconstruction unit 16. To do. The control unit 18 is connected to each element in the display device. The operation of each element of the display device is performed according to the autonomous operation of each component described above or according to an instruction from the control unit 18.

  As described above, in the image display device 1 according to the present embodiment, the pixel position switching unit 15 appropriately switches and outputs the pixel position vectors Xi and Yi for each frame based on the motion vector distribution obtained by the histogram count unit 12. Accordingly, the subfield reconstruction unit 16 is characterized by reconstructing a subfield of a pixel to be reconstructed by switching the pixel position vector Xi or Yi for each frame.

Hereinafter, the configuration and operation of each unit will be described in detail.
FIG. 2 is a diagram for explaining an example of a motion vector histogram created by the histogram count unit 12.

  (A) is an example of a display screen, and it is assumed that the object B is moving in the horizontal direction with respect to the substantially stationary background A. (B) is a histogram of motion vectors in the screen obtained for such a screen. The horizontal axis represents the horizontal component magnitude Vx, and the vertical axis represents the count number Nn. The count number Nn represents the intensity of the bias, and in this figure, a specific motion vector Vx = 0 (count number Nn = 10) and a distribution biased to Vx = 3 (Nn = 20) are shown. Here, it goes without saying that Vx = 0 corresponds to the image of the stationary background A, and Vx = 3 corresponds to the image of the moving object B. The maximum count number Nm = 20, and the motion vector Vm = 3 at that time. Further, as the size of the object B increases, the count number Nn of Vx = 3 increases, and the bias intensity increases. In other words, the size of the object relative to the screen appears as a bias intensity. Therefore, a threshold value S (104) of the count value Nn is provided as an index for evaluating the bias intensity.

  The histogram of (c) is an example in which the bias is large, and shows a case where an object in the screen or the entire screen moves uniformly in a substantially constant direction. On the other hand, the histogram of (d) is an example in which the bias is small, and shows a case where the motion of the object on the entire screen is not uniform.

  The motion vector value V used in the present embodiment is the motion vector value between the reconstruction target frame and the frame preceding the target frame in time. A motion vector whose end point is the reconstruction target pixel is used. In addition, when the result of calculating the pixel position vector is decimal precision, a pixel position vector having integer precision by rounding off, rounding down, rounding up, or the like may be used. Moreover, you may use it with decimal precision. In the following embodiments, the numbers are rounded to integers.

  FIG. 3 is a diagram illustrating an internal configuration of the pixel position switching unit 15. The pixel position vector Xi calculation unit 151 inputs a motion vector V (101) that ends with the reconstruction target pixel of the target frame among the motion vectors detected by the motion vector detection unit 11, and the number and number of subfields. The pixel position vector Xi is calculated using an arithmetic expression. The pixel position vector Yi calculation unit 152 uses the luminance information 103 calculated by the luminance information calculation unit 14 to determine the luminance difference between the pixel indicated by the calculated pixel position vector Xi and the reconfiguration target pixel, and the luminance difference is a threshold value. If it is larger, a new pixel position vector Yi is calculated by correcting the calculated pixel position vector Xi so as to be close to the reconstruction target pixel until a pixel having a luminance difference equal to or smaller than a threshold value appears.

  The motion vector distribution determination unit 153 switches and outputs the pixel position vector Xi or Yi for each frame based on the motion vector distribution (histogram information) 102 obtained by the histogram count unit 12. At this time, a threshold value S (104) for the count number Nn and a threshold value T (105) for the number (kind) of motion vectors W (number) for which the count number Nn is equal to or greater than the threshold value S are given as threshold values for determination. The histogram data 102 input from the histogram count unit 12 is analyzed, and the number W of motion vectors in which the count number Nn is equal to or greater than the threshold value S is obtained. If the bias of the motion vector is large, the value of W is small. Furthermore, the number W of motion vectors is compared with a threshold value T. When W ≦ T, the pixel position vector Xi is output, and when W> T, the pixel position vector Yi is output.

  For example, in the case of FIG. 2C, if S = 10, W = 2, and if T = 3, W ≦ T, so the pixel position vector Xi is output. On the other hand, in the case of FIG. 2D, since W = 5 and W> T, the pixel position vector Yi is output. As described above, when the bias of the motion vector in the screen is large, the pixel position vector Xi obtained by the arithmetic expression is output. When the bias is small, the pixel position corrected so that the luminance difference is equal to or less than the threshold value. Vector Yi is used.

  Here, the threshold value S is appropriately set based on human visual characteristics. For example, the area ratio is preferably set to 20 to 50% of the region. Also, when there are multiple moving objects and their velocities are different, or when there is a single object but the speed is distributed, the moving speed is averaged and the entire object is moving at the average speed. In view of this, the maximum count value Nm and the specific motion vector Vm may be obtained. The threshold value T is the number of areas that occupy a large proportion in the video having the count number Nn equal to or greater than the threshold value S, and is preferably set to 2 to 6, for example.

  FIG. 4 is a diagram illustrating a flowchart of outputting the pixel position vector by switching the pixel position switching unit 15. The following processing is performed for each frame.

In step 111, the pixel position vector Xi (x, y) of each subfield i (i = 1 to N) is obtained using the motion vector V detected by the motion vector detection unit 11. Xi (x, y) is a pixel position vector of each subfield before reconstruction, which is an acquisition destination, based on the pixel position (x, y) to be reconstructed, and is calculated by Expression (1). . Hereinafter, the pixel position vector is simply described as Xi.
Xi = −V × (i−C) / N (1)
Here, i is the number of the subfield to be reconstructed, C is an arbitrary value between 1 and N, and N is the number of subfields constituting one frame. Note that C is a number of a subfield (fixed subfield) in which the light emission data output from the subfield conversion unit 13 is used as it is as a subfield of the reconstruction target pixel, and C is a value between 1 and N. As a result, various subfields can be reconfigured based on the subfield C.

In step 112, the histogram data acquired from the histogram count unit 12 is analyzed, and the number W of motion vectors in which the count number Nn is equal to or greater than the threshold value S is obtained.
In step 113, the number W of motion vectors is compared with the threshold value T. As a result of the comparison, if W ≦ T, the process proceeds to step 114, and the pixel position vector Xi obtained in step 111 is output. As a result of the comparison, if W> T, the process proceeds to step 115.

  In step 115, subfield number N having the largest weight is assigned to i. In step 116, i which is the start of comparison is substituted for j.

  In step 117, the luminance information is acquired from the luminance information calculation unit 14, and the luminance difference between the pixel indicated by the pixel position vector Xi (ie, Xj) obtained in step 111 and the reconstruction target pixel is compared with a threshold value. Here, as the threshold used for the luminance difference determination, for example, the luminance difference in 256 gradation display is preferably set to about 20.

  As a result of the determination, if luminance difference ≦ threshold, the process proceeds to step 118, and the pixel position vector Xj obtained in step 111 is substituted into the pixel position vector Yi. If luminance difference> threshold, the process proceeds to step 119, and the obtained pixel position vector Xj is corrected.

  In step 119, it is determined whether j> C. If j> C, 1 is subtracted from j in step 120. If j> C is not satisfied in step 119, 1 is added to j in step 121. In this subtraction or addition, both are corrected so that j approaches the fixed subfield number C side. As a result, the pixel position vector Xj corresponding to the corrected j is changed to a value representing a position coordinate close to the fixed subfield number C side (that is, a small vector value).

  After changing j, the process returns to step 117 to determine the luminance difference between the pixel indicated by the corrected pixel position vector Xj and the reconfiguration target pixel. Then, the processing from step 119 to step 121 is repeated until the condition of luminance difference ≦ threshold is satisfied. This correction process gradually brings the pixel position vector Xj closer to the reconstruction target pixel until a pixel whose luminance difference is equal to or smaller than the threshold appears. If the luminance difference ≦ the threshold value, the corrected pixel position vector Xj is substituted into the pixel position vector Yi in step 118.

  In step 122, the subfield i is updated, and the process returns to step 116 to obtain the pixel position vector Yi for the next subfield i. If it is confirmed in step 123 that pixel position vectors Yi have been obtained for all subfields i, pixel position vectors Yi are output in step 124.

  As described above, the motion vector distribution is referred to, and the pixel position vectors Xi and Yi are switched and output according to the magnitude of the motion vector bias in the screen.

  In the flowchart of FIG. 4, the pixel position vector Yi is obtained from the pixel position vectors Xi obtained from the motion vectors. However, any other method may be used as long as the pixel position vector is corrected so as to approach the reconstruction target pixel. It may be a method.

  Next, specific examples of subfield reconstruction in the present embodiment are shown separately in FIGS. 5 and 6 according to the motion vector distribution.

  FIG. 5 is a diagram illustrating an example of subfield reconstruction when the distribution of motion vectors is large (uniform). In this case, as shown in FIG. 2C, the number W of motion vectors that are equal to or greater than the threshold value S is small, and the object on the screen or the entire screen moves uniformly in a certain direction. In this case, the pixel position switching unit 15 outputs the calculated pixel position vector Xi as it is to reconstruct the subfield.

  FIG. 5A shows the structure of the subfield before reconstruction. In the figure, the horizontal axis represents the horizontal position of the pixel, the vertical axis represents time, and the display data when the number of subfields N = 6 is shown. Here, it is assumed that the start pixel of the motion vector whose end point is, for example, the pixel n, which is a pixel to be reconstructed, is at a position −6 in the horizontal direction as a relative position with respect to the pixel n. That is, the vector value V of the motion vector is +6.

  FIG. 5B shows a reconstruction result for each subfield of the pixel n. In this case, the pixel position of each subfield to be obtained before reconstruction is obtained by the formula (1) with reference to the reconstruction target pixel. C is an arbitrary value from 1 to N, but C = 4 in this embodiment. As a result of the calculation, the pixel position vector Xi is SF1 is 3, SF2 is 2, SF3 is 1, SF4 is 0, SF5 is -1, and SF6 is -2.

  Since W ≦ T in the determination of step 113 in FIG. 4, the process proceeds to step 114 and uses these Xi values as they are. Therefore, as indicated by arrows 5001 to 5006, SF1 is from the pixel (n + 3), SF2 is from the pixel (n + 2), SF3 is from the pixel (n + 1), SF4 is the original pixel n, and SF5 is the pixel (n-1). From SF6, the subfield emission data is obtained from the pixel (n-2). In this way, the light emission data of each subfield of the reconstruction target pixel n is reconstructed.

  FIG. 5C shows the result of reconstruction of light emission data for all reconstruction target pixels (n−2) to (n + 3). Here, it is assumed that the vector values V of motion vectors having the respective pixels on the reconstruction target frame as the end points are the same +6. Similarly to the case of the pixel n, the pixel position vector Xi is calculated using Equation (1) for each subfield of the reconstruction target pixel. Then, the subfields of the pixel (n−2) to the pixel (n + 3) are reconstructed by the subfield of the pixel indicated by the pixel position vector Xi. As a result, a plurality of subfields (subfields indicated by the same pattern in FIG. 5) arranged in the same pixel in the still image are arranged on the line-of-sight path 5010 after the reconstruction of each pixel.

  FIG. 6 is a diagram illustrating an example of subfield reconstruction when the bias of the motion vector distribution is small (not uniform). This is a case where the number W of motion vectors that are equal to or greater than the threshold value S is large and the object in the screen is moving unevenly as shown in FIG. In this case, the pixel position switching unit 15 corrects the calculated pixel position vector Xi to a pixel position vector Yi and outputs it, thereby reconstructing the subfield. At this time, when the luminance difference between the pixels ≦ the threshold value, the pixel position vector Yi has the same value as Xi. In the present embodiment, a description will be given of the reconstruction of each subfield in the case where the luminance difference> the threshold value is included, that is, in a region other than the similar color region (for example, a region near the edge).

  FIG. 6A shows the configuration of the subfield before reconstruction. In this case, luminance difference> threshold between pixels (n-3) and (n-2), and luminance difference ≦ threshold between other pixels. Also in this case, it is assumed that the vector value V of the motion vector whose end point is, for example, the pixel (n−1) that is the pixel to be reconfigured is +6.

  FIG. 6B shows a reconstruction result of each subfield of the pixel (n−1). First, the pixel position vector Xi is calculated for each subfield of the pixel (n−1) using the equation (1). As a result of the calculation, the pixel position vector Xi (i.e., Xj) is SF1 = 3, SF2 = 2, SF3 = 1, SF4 = 0, SF5 = -1, and SF6 = -2.

  Next, in the determination of step 113 in FIG. 4, since W> T, the process proceeds to step 115 and subsequent steps. In step 117, a luminance difference is determined for each subfield. For example, when j = 6, since Xj for SF6 is −2, the luminance difference between the pixels (n−3) and (n−1) is determined. As described above, since the luminance difference between the pixels (n−3) and (n−1) is larger than the threshold value, the process proceeds to step 119, and 1 is subtracted from j in step 120 to change to j = 5. . Next, returning to step 117, since Xj for SF5 is -1, the luminance difference between pixel (n-2) and pixel (n-1) is determined. Since the luminance difference between the pixels (n−2) and (n−1) is equal to or less than the threshold value, the process proceeds to step 118, and the obtained value −1 of the pixel position vector Xj of SF5 is substituted into the pixel position vector Yi. The luminance difference is similarly determined for other SFs. Since SF1 to SF5 satisfy the luminance difference ≦ threshold value, the value of the pixel position vector Xj is substituted for the pixel position vector Yi. In step 124, the obtained pixel position vectors Yi are output.

  Accordingly, the subfield emission data is acquired as indicated by arrows 6001 to 6006 in FIG. Among these, SF6 acquires light emission data not from the pixel (n-3) but from the pixel (n-2) closer to the reconstruction target pixel (n-1). In this way, the light emission data of each subfield of the reconstruction target pixel (n−1) is reconstructed.

  FIG. 6C shows the result of reconstruction of light emission data for all reconstruction target pixels (n−2) to (n + 3). At this time, the vector values V of the motion vectors having the end points of the pixels (n−2) to (n + 3) on the reconstruction target frame are all the same +6. In addition, it is assumed that the vector values V of the motion vectors whose end points are the pixels (n-4) to (n-3) are all the same +4. Similar to the case of the pixel (n−1), the pixel position vector Yi is calculated using the flowchart of FIG. Then, each subfield of the pixel (n−2) to the pixel (n + 3) is reconfigured by the subfield of the pixel indicated by the obtained pixel position vector Yi. As a result, after reconstruction of each pixel in the similar color area, not only is it arranged on the line-of-sight path 6010, but the subfield reconstruction source is only the subfield of the similar color, and subfields of greatly different colors are displayed. Never get. Therefore, false colors are not generated even in regions other than the similar color region, which is a problem of the prior art, and pseudo contour can be suppressed.

  Here, how the conventional problem is solved by the image display of the present embodiment will be described with reference to FIGS.

FIG. 7 is a diagram for explaining a conventional subfield correction method for preventing a moving image pseudo contour for comparison. The horizontal axis represents the horizontal position of the pixel, the vertical axis represents time, and the display data when the number of subfields N = 6 is shown.
FIG. 7A shows the configuration of the subfield before reconstruction. Here, the light emission state transition of the subfield of the pixel (n-2) of the display data will be described.

  FIG. 7B shows a correction result when the display data is moved by 6 pixels in the horizontal direction, that is, the vector value +6, during moving image display. The light emission data of each subfield of the pixel (n-2) is moved and reconstructed as indicated by arrows 7002 to 7006. The light emission subfield that is actually recognized by the retina is a range (line-of-sight path 7010) sandwiched between two oblique lines. According to this correction, when it is assumed that the image is a still image, the light emission position of a plurality of subfields arranged in the same pixel is changed to the light emission position of the subfield of the pixel position in the line-of-sight path. The contour can be corrected.

  FIG. 7C is a diagram for explaining that a subfield is not reset in some pixels when a motion vector is not uniform, as a problem of a conventional subfield correction method. Here, the correction result by the conventional method when the pixels (n-4) to (n-3) are moved by 4 pixels in the horizontal direction and the pixels (n-2) to (n + 3) are moved by 6 pixels in the horizontal direction is shown. . In the figure, a portion where the subfield is not reset (a portion where the subfield does not emit light) is generated, as in a region 7011 of a frame line.

  As described above, according to the conventional correction method, when the motion of the object in the screen is not uniform, pixels in which the subfield is not set are generated, and the image quality is deteriorated. That is, the luminance of the pixel changes greatly, and a false color composed of pixels having different luminance that are not present in the image is generated.

  On the other hand, in the display method of the present embodiment, a motion vector whose end point is the pixel to be reconstructed is obtained, and resetting is performed for each subfield. Thereby, it is possible to prevent the occurrence of a pixel in which the subfield is not reset.

  8 and 9 are diagrams for explaining the effect of the subfield correction method according to this embodiment.

FIG. 8 shows a method of correcting according to the pixel position vector Xi of this embodiment when the motion vector of each pixel is uniform.
FIG. 8A shows the configuration of the subfield before reconstruction. The motion vector value is +6 for all pixels. Further, the luminance difference> threshold value is set between the pixel (n-3) and the pixel (n-2).

  FIG. 8B is a correction example of the subfield of the pixel (n−1) based on the pixel position vector Xi of this embodiment. The pixel position vector Xi of each subfield of each reconstruction target pixel is calculated using Equation (1). Since the motion vector is uniform, correction based on the luminance difference between pixels is not performed. Then, according to the pixel position vector Xi, as shown by arrows 8001 to 8006, light emission data of each subfield is acquired and reconstructed.

  FIG. 8C shows a result of reconstructing subfields for all pixels using the pixel position vector Xi. At this time, all the subfields in the frame line area 8011 are also reset. Furthermore, as shown in FIG. 8C, the motion vector of each pixel is uniform, and since the subfield of each pixel is corrected with the same correction amount, no false color is generated, and motion blur and pseudo contour are suppressed. It becomes possible to do.

FIG. 9 shows a method of correcting according to the pixel position vector Yi of this embodiment when the motion vector of each pixel is not uniform.
FIG. 9A shows the configuration of the subfield before reconstruction. The motion vector is +4 in pixels (n-4) to (n-3) and +6 in pixels (n-2) to (n + 3). Further, the luminance difference> threshold between the pixel (n-3) and the pixel (n-2), and the luminance difference ≦ threshold between the other pixels.

  FIG. 9B is a correction example of the subfield of the pixel (n−1) based on the pixel position vector Yi of this embodiment. The pixel position vector Xi of each subfield of each reconstruction target pixel is calculated using Equation (1). Next, according to the flowchart of FIG. 4, a pixel position vector Xj that satisfies the luminance difference ≦ threshold value of the pixel from which the subfield is acquired and the reconstruction target pixel is obtained, and the pixel position vector Xj is substituted into the pixel position vector Yi. In this example, the acquisition position of the subfield SF6 is corrected from the pixel (n-3) to (n-2) as indicated by an arrow 9006.

  FIG. 9C shows the result of reconstructing subfields for all pixels using the pixel position vector Yi. At this time, all the subfields in the frame line area 9011 are reset. Furthermore, as shown in FIG. 9C, in the region where the luminance difference is large, the subfield reconstruction target is only the subfield of the similar color, and the subfield of the greatly different color is not acquired. Even when the movement of the object is not uniform, no false color is generated and the pseudo contour can be suppressed.

  With the above method, one target frame can be reconstructed as one new frame. By repeating the process while changing the target frame, a plurality of new frames are generated and an image is displayed. At this time, the pixel position vector is appropriately switched for each frame based on the motion vector distribution, and the correction amount of the subfield is set using the pixel position vector. That is, by switching and displaying FIG. 8C and FIG. 9C for each frame, generation of false colors can be suppressed, and both moving image blur and false contour can be reduced.

  According to the present embodiment, subfield reconstruction can be realized in consideration of a line-of-sight path based on a motion vector, and generation of moving image blur and moving image pseudo contour can be suppressed. Also, if the motion of the object in the screen is not uniform, the subfield reconstruction target is only the subfield of the similar color, and the subfield of the greatly different color is not acquired. The contour can be suppressed. In addition, when an object in the screen or the entire screen moves in the same direction, the subfield reconstruction target includes subfields other than similar colors, and thus it is possible to suppress both moving image blur and pseudo contour. . Then, by switching the subfield correction method for each frame in accordance with the motion vector distribution, it is possible to suppress the generation of false colors and reduce both moving image blur and pseudo contour.

  FIG. 10 is a block diagram showing an image display apparatus according to the second embodiment of the present invention. The image display apparatus 1 of the present embodiment has a configuration in which a motion vector correction unit 19 is added to the configuration of the first embodiment (FIG. 1).

  The motion vector correction unit 19 performs correction to replace the motion vector V (reference numeral 101) detected by the motion vector detection section 11 with another motion vector V ′ (reference numeral 101 ′) according to the distribution state of the motion vectors in the frame. Do. In this case, whether or not correction is necessary is determined with reference to the histogram information 102 obtained by the histogram count unit 12.

  The pixel position switching unit 15 uses the motion vector 101 ′ corrected by the motion vector correction unit 19 to rearrange the subfields of the target pixel, and uses the motion vector 101 ′ to indicate a pixel position vector indicating the pixel of the subfield before reconstruction. calculate. At that time, the histogram information 102 obtained by the histogram count unit 12 and the luminance information 103 calculated by the luminance information calculation unit 14 are referred to.

  The threshold value S (reference numeral 104) and the threshold value T (reference numeral 105) are threshold values used for determination by the motion vector correction unit 19 and the pixel position switching unit 15. Since the operations of the respective units other than the motion vector correction unit 19 and the pixel position switching unit 15 are the same as those in the first embodiment, description thereof is omitted.

  As described above, in the image display device 1 according to the present embodiment, the motion vector correction unit 19 corrects a motion vector suitable for each image based on the motion vector distribution obtained by the histogram count unit 12, and the pixel position switching unit 15. The pixel position vectors Xi and Yi are obtained using the corrected motion vector in step S1, and the pixel position vector Xi or Yi is output by switching for each frame, and the pixel position vector Xi or Yi is used to reconstruct a subfield of a pixel to be reconstructed.

  FIG. 11 is a diagram illustrating an example of an internal configuration of the motion vector correction unit 19. The motion vector correction unit 19 obtains the motion vector distribution (histogram information) 102 from the histogram count unit 12, compares it with the threshold value S (104) and the threshold value T (105), and detects it with the motion vector detection unit 11. Correction for replacing the motion vector value V (101) is performed.

  The register 191 acquires the histogram information 102 from the histogram count unit 12 and holds the maximum value Nm of the count value Nn. The register 192 holds a value Vm of a specific motion vector that gives the maximum count value Nm.

  The comparison correction unit 193 compares the maximum count value Nm with the threshold value S (104). When the maximum count value Nm is equal to or greater than the threshold value S, a motion vector V (101) other than 0 in the corresponding region (region where the histogram is counted) detected by the motion vector detection unit 11 is used as a specific motion vector. Correction (Vn = Vm) for replacement with Vm is performed. However, a motion vector having a value smaller than the motion vector Vm is not replaced, and a vector in the vicinity L is replaced with Vm even if the values indicated by V1 and V2 in FIG. For the horizontal component Vx and the vertical component Vy of the motion vector value, the neighborhood L of each is preferably 1 or less. When the maximum count value Nm is smaller than the threshold value S, the motion vector value V detected by the motion vector detection unit 11 is not replaced (Vn = V).

  The switching unit 194 uses the histogram information 102 to obtain the number W of motion vectors that are equal to or greater than the threshold value S. Next, the number W of motion vectors is compared with a threshold value T. When W ≦ T, the motion vector Vn corrected by the comparison correction unit 193 is output to the pixel position switching unit 15 as a new motion vector V ′ (reference numeral 101 ′). In the case of W> T, the motion vector V detected by the motion vector detection unit 11 is output as it is as the motion vector V ′ (101 ′).

  That is, when an object on the screen or the entire screen is moving uniformly, all the motion vectors Vn with reduced variations are replaced and output. If the motion of the object in the screen is not uniform, the motion vector V detected by the motion vector detection unit 11 is output as it is.

  For example, in the case of FIG. 2C, the maximum count value Nm = 30, and if the threshold value S is 10, Nm ≧ S. Therefore, the motion vector of the entire screen is expressed as Vn = Vm (Vx = 3 ). Next, the vector number W = 2 (Vx = 2 and 3) satisfying Nn ≧ S, and if the threshold value T is 3, W ≦ T. Accordingly, Vn = Vm (Vx = 3) is output to the pixel position switching unit 15 as the motion vector V ′. Incidentally, in the case of FIG. 2D, since W = 5> T, the motion vector V is output as it is.

  The operation of the pixel position switching unit 15 is the same as that in the first embodiment, and is switched for each frame to output the pixel position vector Xi or Yi. However, Vn or V is input as the motion vector V ′. That is, when W ≦ T, a motion vector V ′ (= Vn) with reduced variation is input from the motion vector correction unit 19 and a pixel position vector Xi is output. When W> T, the motion vector V ′ (= V) is input from the motion vector correction unit 19, Xi is corrected using the luminance information 103, and the pixel position vector Yi is output.

  The subfield reconstruction unit 16 reconstructs the subfield using the pixel position vector Xi or Yi output from the pixel position switching unit 15.

  As described above, in the image display device 1 according to the present embodiment, the motion vector correction unit 19 corrects a motion vector suitable for each video based on the motion vector distribution obtained by the histogram count unit 12. The pixel position switching unit 15 obtains pixel position vectors Xi and Yi using the corrected motion vector, and outputs the pixel position vector Xi or Yi by switching for each frame.

  Accordingly, it is possible to greatly reduce the moving image blur and the moving image pseudo contour with respect to the image in which the object in the screen or the entire screen moves in the same direction. In addition, it is possible to suppress the generation of false colors and reduce the moving image pseudo contour for an image in which the motion of an object in the screen is not uniform. Then, by selecting the pixel position vectors Xi and Yi for each frame and switching the correction method of the subfield, it is possible to improve moving image blur and reduce moving image pseudo contour suitable for the image.

  According to the present embodiment, subfield reconstruction can be realized in consideration of a line-of-sight path based on a motion vector, and generation of moving image blur and moving image pseudo contour can be suppressed. Also, if the motion of the object in the screen is not uniform, the subfield reconstruction target is only the subfield of the similar color, and the subfield of the greatly different color is not acquired. The contour can be suppressed. Also, if the object in the screen or the entire screen moves in the same direction, the subfield reconstruction is performed using the motion vector with reduced variation, so that the correction in the area becomes uniform, so that the motion blur and pseudo contour Both of these can be further reduced. Also, by changing the subfield correction method for each frame according to the distribution of motion vectors, it is possible to suppress the generation of false colors and reduce both motion blur and pseudo contour.

  According to each embodiment of the present invention described above, in any case, the deterioration of image quality is more preferably prevented. The above-described embodiments have the following specific effects. In the first embodiment, by switching the subfield correction method for each frame, correction suitable for the video can be realized, generation of false colors is suppressed, and both moving image blur and moving image pseudo contour are reduced. In the second embodiment, by switching the subfield correction method for each frame and correcting the subfield using a motion vector with suppressed variation, an object on the screen or the entire screen moves in the same direction to some extent. Significantly reduces motion blur and motion picture false contours.

Furthermore, in the above embodiments, the following modifications are possible.
In each of the embodiments described above, an example in which the subfield correction method is switched for each frame using the motion vector distribution (histogram) has been described. For example, a motion vector size, a histogram of luminance values, colors, and the like is used. May be switched.

The motion vector has been described using a one-dimensional value with only horizontal movement as an example, but it may be a two-dimensional value. Although the case where the number of subfields is six has been described, the number of subfields may be other than six.
As the luminance difference determined by the pixel position switching unit 15 of each embodiment, a luminance value difference calculated from video RGB data may be used. Further, the difference between individual data of R, G, and B may be used.

  Further, any embodiment of the drawings, methods, etc. described above can be combined to form an embodiment of the present invention.

1 is a block diagram showing an image display device according to a first embodiment of the present invention. The figure explaining an example of the histogram of a motion vector. The figure which shows the internal structure of the pixel position switching part 15. FIG. The flowchart which calculates | requires the pixel position vector of each subfield. The figure which shows an example of the reconstruction of a subfield (when a motion vector is uniform). The figure which shows an example of the reconstruction of a subfield (when a motion vector is not uniform). The figure explaining the correction method of the conventional subfield. The figure explaining the effect of the correction method of the subfield by a present Example (when a motion vector is uniform). The figure explaining the effect of the correction method of the subfield by a present Example (when a motion vector is not uniform). The block diagram which shows the image display apparatus which concerns on 2nd Example of this invention. The figure which shows an example of an internal structure of the motion vector correction | amendment part 19. FIG. The figure explaining the gradation expression method using a subfield. The figure which shows an example of the generation | occurrence | production mechanism of a moving image pseudo contour.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image display apparatus, 10 ... Input part, 11 ... Motion vector detection part, 12 ... Histogram count part, 13 ... Subfield conversion part, 14 ... Luminance information calculation part, 15 ... Pixel position switching part, 16 ... Subfield re-transmission Configuration unit, 17 ... image display unit, 18 ... control unit, 19 ... motion vector correction unit, 193 ... comparison correction unit, 194 ... switching unit.

Claims (8)

In an image display device that divides one frame of an input image into a plurality of subfield periods and reconstructs light emission data in each period of the plurality of subfield periods according to a motion vector of a corresponding pixel between the frames.
A subfield conversion unit that converts the input image into light emission data of a plurality of subfields;
A motion vector detection unit for detecting a motion vector of a pixel between frames for an input image;
A histogram count unit for obtaining a distribution of the detected motion vector for each frame;
A luminance information calculation unit for calculating luminance information of each pixel from the input image;
Among the detected motion vectors, a motion vector whose end point is the reconstruction target pixel in the reconstruction target frame is selected, and data for reconstructing light emission data from the selected motion vector using a predetermined arithmetic expression is selected. A pixel position switching unit that calculates and outputs a pixel position vector indicating an acquisition destination;
The emission data of the subfield of each pixel in the reconstruction target frame output from the subfield conversion unit is the sub-pixel corresponding to the pixel in the reconstruction target frame indicated by the pixel position vector output from the pixel position switching unit. A subfield reconstruction unit that reconstructs using the light emission data of the field;
An image display unit that displays an image using the light emission data of the subfield output from the subfield reconstruction unit,
The pixel position switching unit corrects and outputs the calculated pixel position vector for each frame according to the motion vector distribution obtained by the histogram count unit and the luminance information calculated by the luminance information calculation unit. A characteristic image display device. The image display device according to claim 1,
The pixel position switching unit outputs the calculated pixel position vector Xi when the motion vector distribution obtained by the histogram counting unit is uniform within the frame, and the motion vector distribution is not uniform within the frame. In this case, referring to the luminance information calculated by the luminance information calculation unit, the calculated pixel position until a pixel whose luminance difference between the pixel indicated by the pixel position vector Xi and the reconstruction target pixel is equal to or smaller than a threshold appears. An image display device that outputs a new pixel position vector Yi by correcting the vector Xi so as to be close to the reconstruction target pixel. The image display device according to claim 1 or 2,
A motion vector correction unit that corrects the motion vector V detected by the motion vector detection unit to another motion vector based on the motion vector distribution obtained by the histogram count unit;
The motion vector correction unit counts each motion vector V in the frame detected by the motion vector detection unit within the frame when the distribution of motion vectors obtained by the histogram count unit is uniform within the frame. An image display device characterized in that a detected motion vector V is used as it is when replaced with a motion vector value Vm other than 0 that maximizes the value Nn and the motion vector distribution is not uniform within the frame. The image display device according to claim 2 or 3,
The pixel position switching unit refers to the motion vector histogram obtained by the histogram count unit, obtains a motion vector number W at which the count number Nn is equal to or greater than the threshold value S, and the motion vector number W is equal to or less than the threshold value T. The image display device is characterized in that when the motion vector distribution is uniform and the motion vector number W is larger than a threshold value T, it is determined that the motion vector distribution is not uniform. In an image display method of dividing one frame of an input image into a plurality of subfield periods and reconstructing light emission data in each period of the plurality of subfield periods according to a motion vector of a corresponding pixel between the frames,
Converting the input image into emission data of a plurality of subfields;
Detecting an inter-frame pixel motion vector for the input image;
Obtaining the detected motion vector distribution for each frame;
Calculating luminance information of each pixel from the input image;
Among the detected motion vectors, a motion vector whose end point is the reconstruction target pixel in the reconstruction target frame is selected, and data for reconstructing light emission data from the selected motion vector using a predetermined arithmetic expression is selected. Calculating a pixel position vector indicating an acquisition destination, correcting the calculated pixel position vector for each frame according to the distribution of the motion vector and the luminance information, and outputting the corrected position vector;
Reconstructing the light emission data of the subfield of each pixel in the reconstruction target frame using the light emission data of the corresponding subfield of the pixel in the reconstruction target frame indicated by the pixel position vector;
Displaying an image using the emission data of the subfield;
An image display method comprising: The image display method according to claim 5,
The step of outputting the pixel position vector outputs the calculated pixel position vector Xi when the motion vector distribution is uniform within the frame, and when the motion vector distribution is not uniform within the frame, With reference to the luminance information, the calculated pixel position vector Xi is brought close to the reconstruction target pixel until a pixel whose luminance difference between the pixel indicated by the pixel position vector Xi and the reconstruction target pixel is equal to or smaller than a threshold value appears. And a new pixel position vector Yi is output after the correction. The image display method according to claim 5 or 6,
Correcting the detected motion vector V to another motion vector based on the motion vector distribution;
In the step of correcting the motion vector, when the motion vector distribution is uniform in the frame, each motion vector V in the detected frame is converted into a motion vector other than 0 that gives the maximum count value Nn in the frame. An image display method characterized by using the detected motion vector V as it is when the motion vector distribution is not uniform within the frame. The image display method according to claim 6 or 7,
In the step of outputting the pixel position vector, the motion vector number W is obtained by referring to the motion vector histogram and the count number Nn is equal to or greater than the threshold value S. When the motion vector number W is equal to or less than the threshold value T, An image display method characterized by determining that the distribution of motion vectors is not uniform when the distribution is uniform and the number of motion vectors W is greater than a threshold value T.

Images