JPH09274473A - Device and method for picture processing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the quality of a digital intermediate processing display by detecting the attributes of each picture element data of a picture and controlling the thresholds to quantize the data, based on the attributes. SOLUTION: The R/G/B signals, which form the display picture formed by a plurality of frames, are converted to eight bit data by an A/D converter 101, inputted to a hysterisis error spreading processing circuit 103, converted to two bit data and stored in a frame memory 104. The eight bit data are also inputted into a picture attribute detecting circuit 105, the attributes are detected and the results are stored in a picture attribute information memory 106. The circuit 103 obtains the attributes of the processing object picture elements and the output values of previous frames from the memories 104 and 106, controls the thresholds based on the above data and converts the data into two bit data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method for converting digital image data into image data having a smaller number of bits.

[0002]

2. Description of the Related Art In a display device using a ferroelectric liquid crystal (hereinafter referred to as FLC), the liquid crystal itself retains the display state once set by the application of an electric field even after the application of the electric field is finished. It has a memory property. Therefore, even if the number of scanning lines increases, the contrast itself does not decrease, and it is attracting attention as a device that can realize a large-screen and high-definition display even with a simple matrix configuration.

However, FLC is ON / O in principle.
Since only two states of FF can be taken, it is not possible to express halftone with only one cell in the simple matrix configuration. Therefore, a so-called area gray scale method is adopted in which one pixel is formed by using a plurality of cells, and a gray scale is expressed by a combination of ON / OFF of each cell. In this case, if the area ratio of each cell is a power of 2, such as 1: 2: 4 :, a uniform luminance distribution can be obtained. However, in order to realize more gradation expression by this area gradation, many cell divisions are required, and in particular, it is necessary to make the size of the minimum cell smaller as the number of gradations increases. Is prone to difficulties.

Further, conventionally, a digital halftone processing technique such as a dither method or an error diffusion method has been used as a method for improving the output quality of an output device having a small number of gradations. These are widely used, for example, in printers, facsimiles, digital copiers, etc., but even in the case of a display, the image data of each frame should be halftone processed like a still image and then displayed sequentially. You can

Among the dither methods, the systematic dither method has a simple algorithm, but when the number of colors of the output device is small and the resolution is low as compared with a printer such as an FLC display, the frequency peculiar to the dither pattern is set. It has a property of being newly generated and giving a feeling of noise particularly in a moving image of a natural image. On the other hand, the random dither method is a method in which a random number is generated at each pixel position and used as a threshold value at that pixel position instead of using a constant dither pattern like the systematic dither method. It is said that the method is difficult to put into practical use because the image becomes noisy.

On the other hand, the error diffusion method has a slightly more complicated algorithm than the dither method, but since the image is adaptively processed and the frequency characteristic does not have a peculiarity, it can be used for a computer such as a character / line drawing. Relatively good results can be obtained with both graphics images and natural images.

Here, the error diffusion algorithm will be briefly described. FIG. 16 is a diagram showing an example of a general error diffusion table. FIG. 17 is a diagram for explaining the influence of error diffusion on the pixel of interest.

In the usual error diffusion method, input image data is compared with a threshold value in the order of raster scanning, and converted into data having the number of bits corresponding to the number of gradations that can be displayed by the output device. However, when comparing with the threshold value, the input data is corrected in advance in consideration of the error generated in the pixel processed before that. More specifically, as an example, the error distribution (difference between the input value and the converted value) generated at a certain pixel position is determined by the weight distribution specified by the error diffusion matrix as shown in FIG. Spread to.
On the contrary, when viewed from a certain pixel position, the error is diffused from the processed pixel position as shown in FIG. These diffusion errors are accumulated and added to the input data as a correction component. The comparison with the threshold value is performed using the corrected input data except for the first pixel position.

However, when the error diffusion method is applied to the moving image display, the following problems occur. In the error diffusion method, a change in input data at a certain pixel position results in a change in the value of an error that occurs, and further affects the processing result at an unprocessed pixel position. Therefore, for example, when the mouse cursor moves, the display value of each pixel changes frequently to the right and below the mouse cursor, and the image appears as flicker noise in the eyes of the user.

In addition, the input data originally transmitted as an analog signal from the computer main body is A / D.
If it is obtained by conversion, noise mixed in the transmission system is included in the input data of the error diffusion processing. In that case, even if there is no change in the data in the frame buffer on the computer side, a slight vibration occurs in the input data of the error diffusion processing, which often results in a difference in the error diffusion processing result for each frame. Occurs, and a continuous change in the brightness level of each pixel occurs.

These phenomena are called sparkling noise because they are similar in phenomenon, although the mechanism of occurrence and the place of occurrence are different.

As a measure against the problem of sparkling noise, a method has been proposed in which the threshold value of the error diffusion processing has a kind of hysteresis (hereinafter referred to as a hysteresis error diffusion method). The normal error diffusion method uses a fixed threshold value, but the hysteresis error diffusion method is characterized in that the threshold value is manipulated using the error diffusion result of the previous frame. That is, as a processing procedure, a threshold value to be compared with the input value is first obtained, and then the threshold value is changed so as to be away from the error diffusion result of the previous frame (the threshold value is shifted by a predetermined amount). Then, the output value is determined by comparing the changed threshold value with the input value.

The hysteresis error diffusion method has the effect of widening the range of input values in which the error diffusion processing result at the pixel position currently being processed is the same value as the corresponding pixel in the previous frame by shifting the threshold value. . Therefore, if the threshold shift amount is set to a certain amount, even if the error value diffused at each pixel position changes, in the area where the pixel value does not change in the frame buffer in the computer, The error diffusion processing result hardly changes from the previous frame.

[0014]

However, the above-mentioned theteresis error diffusion method has a problem that the image quality is deteriorated by moving the threshold value. This is particularly likely to occur when a TV image or a scanned image of a photograph is captured and displayed. In the case of a television image, "trailing" occurs in the movement of an object with a gradual brightness gradient because it is affected by the previous frame. When the screen is updated almost constantly like a TV image,
There is no need to apply hysteresis. In addition, if a scanned image of a photograph has a strong hysteresis characteristic, in the case of dither, the influence of the previous frame when first displayed is fixed, and the unnaturalness may not be resolved. . On the other hand, in the case of error diffusion, the value of the error propagated from the processed pixels (upper line and left pixel) changes due to noise and image movement, but strong hysteresis characteristics should be provided. When trying to resist with, a part that can be completely resisted and a part that cannot be completely mixed are mixed, and in the case of a natural image, it becomes an unnatural image. In such a case, it is preferable not to have the hysteresis characteristic even if the sparkling noise is allowed.

The present invention has been made in view of the above problems, and it is possible to detect the attribute of each pixel data in an image and control the threshold value for quantizing the pixel data based on the attribute. It is an object of the present invention to provide an image processing method and apparatus that improve the quality of digital halftone processing display under a noise condition or the like when a plurality of frames of images are sequentially displayed to form a display image.

[0016]

The image processing apparatus of the present invention for achieving the above object has the following configuration. That is, an image processing device for converting first image data having a first gradation number forming a display image formed by a plurality of frames into second image data having a second gradation number, A conversion unit that converts each pixel value of the first image data into a pixel value of the second image data based on a set threshold value, and holds the pixel value of the second image data obtained by the conversion unit. Holding means, detecting means for detecting the attribute of each pixel in the first image data, pixel value of the previous frame corresponding to the pixel to be processed held by the holding means, and the detection Control means for controlling the threshold value for the pixel in the conversion means based on the attribute of the pixel detected by the means.

Further, preferably, the conversion means converts each pixel value of the first image data into a second pixel value by an error diffusion process.
The pixel value in the image data, the control means is used in the error diffusion processing based on the pixel value of the previous frame corresponding to the pixel to be processed and the attribute of the pixel, Control the threshold applied to the pixel. This is because when the hysteresis error diffusion process is applied, the adverse effect on the image quality can be reduced.

Further, preferably, the conversion means converts each pixel value of the first image data into a pixel value in the second image data by dither processing, and the control means
The threshold value applied to the pixel in the dither matrix in the dither processing is controlled based on the pixel value of the previous frame corresponding to the pixel to be processed and the attribute of the pixel.

Further, preferably, the detecting means detects the attribute in units of a group including a plurality of pixels.

Further, preferably, the group is a block of M × N pixels obtained by dividing an image of one frame into p × q.

Further, preferably, the detecting means detects the attribute of each group based on the characteristics of the image in each group.

Further, preferably, the detecting means detects the attribute of each group based on whether or not the image in each group has changed from the image of the previous frame.

Further, preferably, the control means uses a predetermined threshold value when the detection means detects that the image of the group to which the pixel to be processed belongs is changed, and uses the predetermined threshold value. If it is detected that the threshold value has not changed, the threshold value is changed to increase the threshold value interval for the output value of the pixel in the previous frame. This is because by changing the threshold value, it is possible to eliminate the adverse effect on the moving image portion.

Further, preferably, the detecting means detects the attribute of each group based on the frequency characteristic of the image in each group.

Preferably, the detecting means detects the attribute of each group based on whether or not there is a peak in at least one of the low frequency component and the high frequency component in the frequency characteristic of the image of each group.

Further, preferably, the control means, when the detection means detects a peak in at least one of a low frequency component and a high frequency component in a frequency characteristic of an image of a group to which the pixel to be processed belongs, the pixel is detected. The threshold value is changed so as to increase the interval of the threshold value with respect to the output value in the preceding frame, and a predetermined threshold value is applied in other cases. For example, in the display area of computer graphics or the like, generation of sparkling noise or the like is prevented by changing the threshold value, and in the display area of natural images or the like, image deterioration due to the change of the threshold value is prevented.

Further, in the image processing method of the present invention which achieves the above object, the first image data having the first gradation number forming the image for the display formed by a plurality of frames is converted into the second gradation. An image processing method for converting into second image data having a number, the converting step of converting each pixel value of the first image data into a pixel value in the second image data based on a set threshold value. A holding step of holding the pixel value of the second image data obtained in the converting step, and a detecting step of detecting the attribute of each pixel in the first image data,
The threshold for the pixel in the conversion step is based on the pixel value of the previous frame corresponding to the pixel to be processed held in the holding step and the attribute of the pixel detected in the detection step. And a control step for controlling the value.

[0028]

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

[First Embodiment] FIG. 1 is a block diagram showing the arrangement of a display interface system according to the present embodiment. In this system,
Framed analog moving image signals of red (R), green (G), and blue (B) are input, and this is converted by an A / D converter 101 into a multi-value digital signal of 8 bits for each color.
Reference numeral 102 denotes a PLL circuit, which inputs a synchronizing signal in the image signal to generate a dot clock and supplies the dot clock to the A / D converter 101. The A / D converter 101 samples the analog moving image signal input in synchronization with the dot clock and outputs the multilevel digital signal.

The hysteresis error diffusion processing circuit 103 is
A multi-value digital signal from the A / D converter 101 is input, hysteresis error diffusion processing is performed, and the result is stored in the frame memory 104 (the algorithm of the hysteresis error diffusion processing and a block diagram of the circuit configuration will be described later). .

The image attribute detection circuit 105 reads the multi-valued digital signal from the A / D converter 101, checks the image attribute at each pixel position, and writes the result in the image attribute information memory 106 (specific image attribute information). The contents will be described later). The hysteresis error diffusion processing circuit 103 reads the attribute information of the pixel to be processed from the image attribute information memory 106, and uses the information to determine the threshold value used for the hysteresis error diffusion processing.

On the other hand, the motion detection circuit 108 receives the multilevel digital signal from the A / D converter 101, detects a changed line from the image of the previous frame, and stores the information in the partial rewriting line information memory. Write to 109.

It is known that the FLCD 110 has a property (so-called memory property) that the display state is maintained even after removing the electric field after applying the electric field for driving scanning to each display cell. Utilizing this property, CRT
It is possible to achieve both quick response to image changes and stable display by preferentially scanning the lines in which changes are detected, rather than the regular scanning as seen in driving other devices such as displays. it can. Therefore,
Based on the information written in the partial rewriting line information memory 109, the memory / scanning control circuit 107 sets the F line so that the line changed from the previous frame is preferentially scanned.
It controls the order of lines for transferring data to the LCD 110. The FLCD 110 displays the received display data on the designated line.

With the above configuration, the hysteresis error diffusion circuit 103 of this embodiment acquires the image attribute information of the pixel to be processed from the image attribute information memory 106, and the error of the pixel concerned is acquired based on the image attribute information. Controls the threshold applied to the diffusion process. In the first embodiment, a case where a change in image signal at each pixel position is used as image attribute information will be described.

FIG. 2 is a block diagram showing the configuration of a display interface system using the change of the image signal at each pixel position as the image attribute information. 2, configurations having the same functions as the configurations shown in FIG. 1 are assigned the same numbers as in FIG. In this example, since the change of the image signal is used as the image attribute information, the image attribute detection circuit 105 detects the change of the image signal.
Therefore, the image attribute detection circuit 105 in FIG. 1 is also used as the motion detection circuit 108. Further, the image attribute information memory 106 stores information indicating a change in the image, which is used as the motion information memory 306 in FIG.

That is, in the configuration of FIG. 2, the result of the motion detection circuit 305 detecting the change in the image signal at each pixel position is written in the motion information memory 306, and the hysteresis error diffusion processing circuit 103 processes each pixel data. Before that, the motion information is read out and is used as pixel attribute information for processing.

Here, the algorithm of the hysteresis error diffusion processing will be described. FIG. 3 is a diagram showing threshold values normally used when the output signal has four values (two bits). In the simple error diffusion processing, when the input signal value corrected by the diffusion error is other than the four values that can be output, the input signal has a threshold value between the two output values that sandwich the value. The two values are compared and one of the two values is selected as the output signal value.

For example, assume that the corrected input signal value is 110. In this case, two values that can be output (85, 1,
The threshold between 70) is 128 and the threshold to be compared with the input value is 128. Where 110 <1
Since it is 28, the output signal value is 85, and the error value to be diffused is +25.

On the other hand, in the hysteresis error diffusion processing, the threshold value is operated according to the output result of the previous frame stored in the frame buffer. FIG. 4 is a diagram showing an example of a threshold value operation in the hysteresis error diffusion processing.

The threshold operation policy is that when the output signal value of the previous frame is smaller than the corrected input signal value of this time,
If the output value of the previous frame is larger than the corrected input signal value of this time on the contrary, the threshold value is changed upward (for example, +20 is changed), and the threshold value is changed downward (for example, −20 change). To let). As shown in FIG. 4, when the corrected input signal value is 110 and the output signal value of the previous frame is 85, the threshold value is changed to 148. Therefore, the output signal value becomes 85. On the other hand, even if the corrected input signal value is 110, if the output signal value of the previous frame is 170, the threshold value is changed to 108, so that the output signal value is 17
It becomes 0.

FIG. 5 is a block diagram showing the configuration of the hysteresis error diffusion circuit 103. The input original input signal value is input to the input value correction circuit 601. The input value correction circuit 601 reads the error diffused / accumulated at the pixel position currently being processed stored in the diffusion error memory 605, calculates the sum of this error value and the original input signal value, and corrects the input signal value. Output as. The input / threshold comparison circuit 602 is
Data of the pixel position currently being processed (data of the previous frame) is read from the frame memory 104, and an appropriate threshold value is selected and read from the threshold value memory 603 according to the size of the data and the correction input signal value.

The output value is read from the quantization memory 604 and written in the frame memory 104 based on the result of comparison between the threshold value read from the threshold value memory 603 and the corrected input signal value. On the other hand, the error calculation circuit 607 calculates the error value of the output signal value with respect to the corrected input signal value,
Output to the diffusion error calculation circuit 606. The diffusion error calculation circuit 606 calculates the value of the error diffused to the peripheral pixels by using a predetermined error diffusion matrix (for example, the error diffusion matrix as shown in FIG. 16) and stores the obtained value in the diffusion error memory 605. Write in the corresponding position.

6 and 7 show the input / threshold comparison circuit 6
It is a flowchart which shows the operation algorithm of 02.
In the figure, q (i) (i = 0, 1, 2, 3) is a value that can be output and represents 0, 85, 170, 255 in the order of the size of i.

In steps S801, S803, and S805, the input / threshold comparison circuit 602 sequentially compares the correction input signal value with q (i), and the correction input signal value is equal to or equal to the output possible value. If there is not, it is checked which output possible value is the closest and larger, and the value of i of the output possible value q (i) is obtained (steps S802, S8).
04, S806, S807).

Next, in step S901 (FIG. 7), the motion information of the group to which the pixel currently being processed belongs is read from the motion information memory 306, and the output value at the same pixel position of the previous frame in the frame memory 304 Either Th, Th + or Th− corresponding to the value of i is selected from the threshold table. FIG. 8 is a diagram showing a threshold table in the present embodiment. The threshold value table stores a standard threshold value Th, an upper threshold value Th +, and a lower threshold value Th− corresponding to each index (i = 0 to 3). In addition, Th, Th +
Alternatively, how to select Th− will be described later. Finally, the corrected input signal value is compared with the selected threshold value, and either q (i) or q (i + 1) is output according to the result.

Next, the motion detection circuit 108 in FIG. 2 will be described. Motion detection circuit 30 in this embodiment
Reference numeral 5 detects the presence / absence of motion for each group including a predetermined number of pixels.

FIG. 9 is a diagram showing how the entire screen of the display is divided into groups of 8 × 8 pixels. The motion detection circuit 108 moves (changes) the pixel data between frames.
Are detected in units of this group, and the detection method is based on the dither method.

FIG. 10 is a block diagram showing the configuration of the motion detection circuit of this embodiment. The input / dither threshold comparison circuit 1101 receives the original input signal of R / G / B, and selects the threshold stored in the dither threshold memory 1102 according to the position of the pixel being processed. take in. Then, the original input signal is compared with the threshold value, and the dither data obtained as a result is output. Since the result of this dither processing is not used for display, it is not necessary to use FLC.
It is not necessary to perform the dithering process for the number of bits corresponding to the number of display colors of D110. Therefore, in this example, in order to simplify the circuit, binary dither processing is performed for each color.

The output result of the input / dither threshold value comparison circuit 1102 (current dither, referred to as data) is subsequently input to the dither / data comparison circuit 1103. The dither data comparison circuit 1103 reads dither data (referred to as previous dither data) corresponding to the same position in the previous frame from the dither data memory 1104 and compares it with the current dither data from the input / dither threshold comparison circuit 1101. Compare. If the current dither data and the previous dither data are the same in all R, G, and B as a result of the comparison,
In the motion information memory 306, "1" is set to R, G, at the storage position corresponding to the group to which the pixel being processed belongs.
If any one of B is different, "0" is ORed with the already written value and written. However, in the case of the first pixel of the group, "1" or "
0 "is overwritten as it is. In this way, the motion information memory 306 stores the motion information in units of groups (0 when there is motion and 1 when there is no motion).

FIG. 11 is a block diagram showing the structure of the motion information memory in this embodiment. As shown in FIG. 11, the inside of the motion information memory 306 includes two banks A1203 and B12 each having a memory of the same capacity.
It has a double buffer structure of 04. Then, the writing from the motion detection circuit 305 and the reading from the hysteresis error diffusion processing circuit 303 are controlled so that they are alternately switched in frame units while being kept exclusive. That is, for example, bank A1203
On the other hand, in the frame period in which the motion information is stored by the motion detection circuit 305, the motion information of the previous frame is read from the bank B 1204 as image attribute information by the hysteresis error diffusion processing circuit 103.

The hysteresis error diffusion processing circuit selects the threshold value by using the motion information in the previous frame of the group to which the pixel currently processed is stored, which is stored in the motion information memory 306 as described above (step). S90
1), the processing procedure will be described below.

FIG. 12 is a flow chart for explaining the threshold value selection procedure in this embodiment. Step S1
In 201, the input / threshold comparison circuit 602 acquires the motion information of the group to which the current pixel to be processed belongs from the motion information memory 306. In step S1202, it is determined whether the motion information of the pixel group is “0 (moving)” or “1 (no moving)”. Step S
If it is determined in 1202 that there is movement, the process advances to step S1203. In step S1203, the standard threshold value Th (see FIG. 8) of the index i is read from the threshold value memory 603 and set as the threshold value of the error diffusion process of the pixel. That is, when the value of the motion information is "0", the threshold value Th is selected regardless of the output signal value of the previous frame.

On the other hand, if it is determined in step S1202 that there is no motion, the flow advances to step S1204 to obtain the output value corresponding to the pixel of the previous frame from the frame memory 104. Then, in step S1205, the standard threshold Th of the index i and step S120
The output value of the previous frame acquired in 1204 is compared.
If the output value of the previous frame is smaller than the standard threshold value of index i, step S1206
From step S1207, the upper threshold value Th + of the index i is selected and set as the threshold value of the error diffusion processing of the pixel. On the other hand, if the output value of the previous frame is equal to or larger than the standard threshold value of index i, step S1206
From step S1208, the lower threshold value Th− of the index i is selected and set as the threshold value of the error diffusion process of the pixel.

As described above, when the value of the motion information is "1", the output signal value of the previous frame is compared with Th, and Th
When Th is smaller, Th + is selected, and when Th is smaller, Th- is selected.

By the way, the dither data comparison circuit 110
The comparison result in 3 is the partial rewriting line information memory 109.
Also written in. FIG. 13 is a diagram showing a concept of processing when a motion is detected. As shown in the figure, the partial rewriting line information memory 109 has a 1-bit flag (partial rewriting flag 1302) corresponding to each line. As a result of the comparison in the dither data comparison circuit 1103, if the current frame and the previous frame have different dither data, all the flags corresponding to the line including the pixel are set. If both dither data are the same, the flag is not written.

On the other hand, the memory / scanning control circuit 107
Checks these flags in the order determined in order, and when it finds the set flag, immediately instructs the FLCD 309 to rewrite the line, and at the same time resets the corresponding flag. The access to the partial rewrite flag 1302 from the motion detection circuit 108 and the memory / scan control circuit 107 described above is performed asynchronously.

The dither data comparison circuit 1103 writes to the group unit motion information memory 1105 and the partial rewriting line information memory 1106, and then the dither data of the current frame is placed at a predetermined position in the dither data memory 1104.
Write the data.

As described above, according to the first embodiment, the standard threshold value is used for the pixels in the group in which the motion is detected, regardless of the output value of the previous frame, and the motion is detected. For the pixels in the group that did not exist, the error diffusion process is performed by changing the threshold value according to the output value of the previous frame. Therefore, the sparkling noise is reduced and the deterioration of the image quality in the moving image portion is reduced.

[Second Embodiment] As a second embodiment, the image attribute detection circuit 105 of FIG. 1 detects not the movement (change) of the image information but the characteristic of the image, and thereby the hysteresis error diffusion circuit 103 is operated. A configuration for changing the processing will be described.

FIG. 14 is a block diagram showing the internal structure of the image attribute detection circuit 105 according to the second embodiment.
The original input value of R / G / B is once the image data memory 1
It is stored in 401. Image frequency characteristic detection circuit 1402
Performs frequency analysis by DCT (Discrete Cosine Tranform) on the data stored in the image data memory 1401 for each 8 × 8 block group. As a result, when there is a peak in at least one of both ends of the low frequency and the high frequency, "1" is set as the image attribute information of the group, and "0" is set as the image attribute information of the group otherwise. Image attribute information memory 106
Write (overwrite) to the storage location corresponding to the relevant group. This is to extract the features of the image such as characters and line drawings in the former and the feature of the natural image in the latter, but it is also possible to use any method other than this method as long as a similar result can be obtained. .

By the way, the image attribute detection circuit 1404 (corresponding to 105 in FIG. 1) does not necessarily have to detect the image attribute every frame, and may be performed asynchronously with the hysteresis error diffusion process. Therefore, the image data memory 1401 does not need the memory capacity for one screen, and may have only the capacity corresponding to the image area in which the image attribute can be detected as described above within one frame period. Here, 8x
The capacity is one horizontal row of eight groups. That is, it is ideal to examine the image attribute for each frame, but it is generally considered that the image attribute such as a natural image / computer-generated image rarely changes in frame units. Therefore, for example, it is sufficient to check each pixel once in several frames and update the image attribute information memory 106,
By doing so, the amount of memory required for the image data memory 1401 will be sufficient if it can be processed in one frame period. However, this is not the case when the presence or absence of image change is used as the image attribute, but in this case, the configuration is as shown in FIG. 2 and the image data memory 1401 is not used.

The process of selecting the threshold value is similar to that of the first embodiment (FIG. 12). That is, when performing the processing, the hysteresis error diffusion processing circuit 103 reads the image attribute information of the pixel group to which the pixel to be processed belongs from the image attribute information memory 106, and when the value is “1”, the threshold value is set. The error diffusion process is performed using either Th + or Th− using the processing result of the previous frame (steps S1201, S1202, S1204 to S1).
208). However, when the value of the image attribute information is “0”, Th is used as the threshold value regardless of the processing result of the previous frame (step S1203).

As described above, according to the second embodiment, the hysteresis error diffusion process is executed for the area of the computer graphics image such as the character and the line drawing.
The normal error diffusion process is applied to the natural image area. It is possible to reduce sparkling noise that tends to occur on a computer graphics image and display a natural image and the like in high image quality.

[Third Embodiment] In the first and second embodiments described above, the hysteresis error diffusion process is used.
In the third embodiment, hysteresis dither is used instead of the hysteresis error diffusion process. That is, the hysteresis error diffusion circuit 103 in FIG. 1 is replaced with a hysteresis dither circuit.

FIG. 15 is a block diagram showing the structure of the hysteresis dither circuit according to the third embodiment. As can be seen from the figure, generally, the error generated in the narrow sense of dither processing is not diffused to the peripheral pixels, so that the input value correction circuit 601, the error calculation circuit 607, and the diffusion error calculation circuit 606 in the hysteresis error diffusion circuit (FIG. 6). , Diffusion error memory 605 is not necessary. However, other than that, the configuration can be common in the case of the error diffusion process and the case of the dither. That is, in the hysteresis dither processing, the dither processing result at the pixel position of the previous frame is referred to, at the same time, the image attribute of the pixel is referred to, and the threshold value is changed based on both.

Therefore, in the case of hysteresis dither,
In the block diagrams of FIGS. 1 and 2, the hysteresis error diffusion processing circuit 103 may simply be replaced with the dither processing circuit 1700. Further, the threshold value operation may be performed in the same manner as in the error diffusion.

Even if the present invention is applied to a system composed of a plurality of devices (for example, a host computer, an interface device, a reader, a printer, and the like), an apparatus (for example, a copying machine, a facsimile, etc.) comprising one device Device).

An object of the present invention is to provide a storage medium storing a program code of software for realizing the functions of the above-described embodiments to a system or an apparatus, and to provide a computer (or CPU) of the system or apparatus.
And MPU) read and execute the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD
-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

When the computer executes the readout program code, not only the functions of the above-described embodiment are realized, but also the OS (Operating System) running on the computer based on the instruction of the program code. ) May perform some or all of the actual processing, and the processing may realize the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instructions of the program code, It goes without saying that the CPU included in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0073]

As described above, according to the present invention,
It is possible to examine the attribute of the pixel in the image and control the amount of change in the threshold value operation in consideration of hysteresis based on the attribute, and it is possible to suppress the image quality deterioration newly generated by the threshold value operation. Therefore, when the images of a plurality of frames are sequentially displayed to form a display image, the quality of the digital halftone processing display under a noise condition or the like is improved.

[0074]

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a display interface system according to the present embodiment.

FIG. 2 is a block diagram showing a configuration of a display interface system using a change in an image signal at each pixel position as image attribute information.

FIG. 3 is a diagram showing threshold values normally used in a 2-bit error diffusion process.

FIG. 4 is a diagram showing an example of a threshold value operation in hysteresis error diffusion processing.

5 is a block diagram showing a configuration of a hysteresis error diffusion circuit 103. FIG.

FIG. 6 is a flowchart showing an operation algorithm of the input / threshold comparison circuit 602.

FIG. 7 is a flowchart showing an operation algorithm of the input / threshold comparison circuit 602.

FIG. 8 is a diagram showing a threshold table in the present embodiment.

FIG. 9 is a diagram showing a state where the entire screen of the display is divided into groups of 8 × 8 pixels.

FIG. 10 is a block diagram showing a configuration of a motion detection circuit of this embodiment.

FIG. 11 is a block diagram showing a configuration of a motion information memory according to the present embodiment.

FIG. 12 is a flowchart illustrating a threshold selection procedure according to the present embodiment.

FIG. 13 is a diagram showing a concept of processing when a motion is detected.

FIG. 14 is an image attribute detection circuit 10 according to the second embodiment.
5 is a block diagram showing an internal configuration of FIG.

FIG. 15 is a block diagram showing a configuration of a hysteresis dither circuit according to a third embodiment.

FIG. 16 is a diagram showing an example of a general error diffusion table.

FIG. 17 is a diagram for explaining the influence of error diffusion on a pixel of interest.

[Explanation of symbols]

 101 A / D converter 102 PLL 103 Hysteresis error diffusion processing circuit 104 Frame memory 105 Image attribute detection circuit 106 Image attribute memory 107 Memory / scanning control circuit 108 Motion detection circuit 109 Partial rewriting line information memory 110 FLCD

Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location H04N 1/405 G06F 15/68 310A 5/66 H04N 1/40 C

Claims (12)

[Claims] 1. An image processing apparatus for converting first image data having a first gradation number forming a display image formed by a plurality of frames into second image data having a second gradation number. A conversion unit that converts each pixel value of the first image data into a pixel value of the second image data based on a set threshold value; and a pixel of the second image data obtained by the conversion unit. Holding means for holding a value, detecting means for detecting an attribute of each pixel in the first image data, and a pixel value of a previous frame corresponding to the pixel to be processed held in the holding means, An image processing apparatus, comprising: a control unit that controls a threshold value for the pixel in the conversion unit based on the attribute of the pixel detected by the detection unit. 2. The conversion means converts each pixel value of the first image data into a pixel value in the second image data by an error diffusion process, and the control means converts the pixel value into the pixel to be processed. The image processing according to claim 1, wherein a threshold value applied to the pixel used in the error diffusion processing is controlled based on a pixel value of a corresponding previous frame and an attribute of the pixel. apparatus. 3. The conversion unit converts each pixel value of the first image data into a pixel value in the second image data by dither processing, and the control unit corresponds to the pixel to be processed. The image processing apparatus according to claim 1, wherein the threshold value applied to the pixel in the dither matrix in the dither processing is controlled based on the pixel value of the previous frame and the attribute of the pixel. . 4. The image processing apparatus according to claim 1, wherein the detection unit detects an attribute in units of a group including a plurality of pixels. 5. The group consists of one frame of image p
The image processing device according to claim 4, wherein the image processing device is a block of M × N pixels obtained by dividing the image into q × q. 6. The image processing apparatus according to claim 4, wherein the detection unit detects the attribute of each group based on the characteristics of the image in each group. 7. The method according to claim 4, wherein the detection unit detects the attribute of each group based on whether the image in each group has changed from the image of the previous frame. Image processing device. 8. The control means uses a predetermined threshold when the detection means detects that the image of the group to which the pixel to be processed belongs is changed, and the image of the group is changed. The image processing apparatus according to claim 7, wherein when it is detected that the threshold value is not present, the threshold value is changed so as to widen the threshold value interval for the output value of the pixel in the previous frame. 9. The image processing apparatus according to claim 4, wherein the detecting unit detects the attribute of each group based on the frequency characteristic of the image in each group. 10. The detecting means detects the attribute of each group based on whether or not a peak exists in at least one of a low frequency component and a high frequency component in the frequency characteristic of the image of each group. The image processing device according to claim 9. 11. The preceding frame of the pixel when the control unit detects a peak in at least one of a low frequency component and a high frequency component in the frequency characteristic of an image of a group to which the pixel to be processed belongs in the detection unit. 11. The image processing apparatus according to claim 10, wherein the threshold value is changed so as to widen the interval of the threshold value with respect to the output value, and the predetermined threshold value is applied in other cases. 12. An image processing method for converting first image data having a first gradation number forming a display image formed of a plurality of frames into second image data having a second gradation number. A conversion step of converting each pixel value of the first image data into a pixel value of the second image data based on a set threshold value; and a pixel of the second image data obtained in the conversion step. A holding step of holding a value, a detecting step of detecting an attribute of each pixel in the first image data, and a pixel value of a previous frame corresponding to the pixel to be processed held in the holding step, And a control step of controlling a threshold value for the pixel in the converting step based on the attribute of the pixel detected in the detecting step.