TW201503057A - Method and device for enhancing partial image contrast of histogram - Google Patents

Abstract

This invention relates to a method and a device for enhancing partial image contrast of a histogram. A histogram distribution of a figure or an image is first obtained, and then a brightness range is selected for compressing the grayscale space of an immaterial brightness region of the histogram, such that the pixels of the immaterial brightness region are squeezed to a small range, while the pixels of a material brightness region can obtain a greater brightness grayscale equalization space. Then, a specific histogram stretching contrast enhancement method (HS) or histogram equalization contrast enhancement method (HS) is used to expand the grayscale range of the material block of the figure for obtaining more obvious contrast enhancement and image clarity, so as to solve the problem of figure content distortion caused by unbalanced distribution of brightness grayscale of pixels in other regions of the figure or image when the brightness grayscale of the whole figure or image is partially enhanced.

Description

Histogram local image contrast enhancement method and device

The present invention is a device for enhancing image contrast, and more particularly to a device for extending the histogram range of a partial image to enhance image contrast.

The contrast of the existing single image (photo) or continuous image (film) plays an important role in distinguishing the quality of the image by the human eye. If the shooting environment is not ideal, the contrast of the obtained image may be insufficient to produce an overall image. In the case of fatigue, in order to solve the problem of insufficient image contrast, various contrast-enhancing algorithms have emerged.

By using the expansion of the contrast of the image, the image can be visually more strongly felt, thereby achieving the purpose of enhancing the sharpness of the image. The common gray-scale histogram (Histogram) function is the number of times the pixels in the reaction image appear in each gray level. As shown in the original image shown in FIG. 1, FIG. 2 is a gray-scale histogram statistical result of the original image of FIG. 1, wherein the horizontal coordinate of the gray-scale histogram is a gray level, and the ordinate is the number of pixels in the gray level. It can be seen from the gray-scale histogram of FIG. 2 that most of the pixels in FIG. 1 are located between the gray levels 20 to 200. Therefore, the gray-scale histogram is marked as the most basic statistical feature of the image. The principle of image contrast enhancement is to make the human eye visually perceive more image detail texture and edge parts by expanding the contrast between adjacent histograms.

The existing image contrast enhancement techniques based on the histogram of the contrast enhancement processing are mainly divided into two categories, namely Histogram StretchingBased Methods and Histogram EqualizationBased Methods. The contrast enhancement technology, its technical content is explained as follows.

Histogram StretchingBased Methods use the original image histogram to re-stretch its new image histogram distribution with its global grayscale range. The most famous method is Histogram Stretching (HS). Improved Histogram Stretching method (IHS) and other techniques. The advantage of this technique is that it only stretches the gray scale distribution of the original histogram without changing the brightness distribution curve of the original image, so the algorithm is simpler and does not change. The difference in brightness of each pixel does not cause image distortion; however, since the histogram of most images has little pixel area in the luminance grayscale value range, there is no pixel distribution. Often, the range of brightness that can be stretched is not found and no effect can be produced.

The HS method uses the grayscale value of the original image as a histogram analysis, and then stretches its histogram to the full dynamic region, as shown in Figures 3 to 5. Figure 3 shows the original histogram, and Figure 4 shows The histogram of FIG. 3 is subtracted from the minimum value (0 to 130), and FIG. 5 is the histogram of FIG. 4 after the reduced histogram is transmitted through the HS method, and an extended histogram is generated. The HS method is suitable for using an image in a histogram set, and a new output pixel value is obtained using Equation (1), wherein the value 255 is the maximum brightness value of the 8-bit input image. [Number 1]

Based on the HS method, Hee-Chul Kim proposed the IHS method, which can be applied to the method of image processing on hardware, which is represented by equation (2): [Number 2] In the IHS method, the intensity of the extended histogram is determined by the weight parameters M(Multiple) and US (User Select) to obtain the pixel value of the output image. The value of the weight parameter M determines the main expansion strength, and the weight parameter US is set by the user. The parameter is mainly to fine-tune the intensity of the weight parameter M. The IHS method improves the original HS to have a division formula, which has the advantage that it is easier to implement on a hardware.

Histogram EqualizationBased Methods are mainly used to nonlinearly stretch the original image histogram and redistribute the image pixel values so that the original image histogram is a balanced histogram distribution. The HE method is more commonly used because it produces a more uniform brightness distribution, so image contrast enhancement is currently widely used. This method is often used to increase the global contrast strength of many images. First, an original image is defined as X, the total number of pixels of the original image is N, and the grayscale range of the image is set to [0, L-1], and L represents the largest grayscale value of the image pixel. The Probability Density Function (PDF) is defined as Equation (3): [Number 3] The above k represents the gray class, Represents the number of pixels that appear on the gray level. Figure 6 shows the statistical results after calculation by PDF. The Cumulative Density Function (CDF) is a progressive calculation based on PDF, as shown in equation (4): [Number 4] The result of CDF statistics, as shown in Figure 7, the maximum value of the progressive result will be equal to 1; finally the transfer function through the image, as shown in equation (5), [5] A histogram equalization conversion process is performed to obtain a contrast-enhanced image. As shown in FIG. 8 , after the HE method, the corresponding new histogram distribution is an extension of the global equilibrium between 0 and L-1.

Due to the direct and simple characteristics of the HE method, the contrast enhancement method based on the HE method has been published in recent years. Many improved techniques have been published in recent years. The BBHE method is used as a representative of one of the technical representatives: First, the BBHE uses the histogram of the original input image. The average luminance value (Xm) is divided into two sub-Histogram regions, and then the histogram equalization processing is performed for the two regions, as shown in FIG. Similar methods are DSIHE and SSTHE. The histogram is also cut by the average brightness or the intermediate brightness, and then the individual histogram equalization is performed on each sub-histogram.

Another method is to consider the correlation between the pixel position and the gray level of its pixel to achieve contrast enhancement. Take the Contrast limited adaptivehistogram equalization method (CLAHE) as an example. The steps are as follows: (1) Partition Block: The input image is divided into non-overlapping sub-blocks of equal size, and the total number of pixels contained in each sub-block is set to M. The larger the range set by the sub-block, the more obvious the effect of contrast enhancement, but it is easy to cause the image details to be easily lost. (2) Calculate the histogram: Represents a histogram of sub-blocks, k represents the gray level, and its gray level is set from 0 to L-1, where L represents the number of gray levels that may appear. (3) Calculate the restriction level : Calculate the Restricted Level using Equation (6), where Maximum slope The setting determines the magnitude of the contrast enhancement, and the value is set in the range of 1 to 4; It is called the limit level coefficient, and its value is set in the range of 0~100. when When it is equal to 0, Minimum value ;when When it is equal to 100, take the maximum value At this time, the corresponding contrast stretching effect is most obvious. [Number 6] (4) Pixel point reallocation: according to the previous step, for each sub-block, use the corresponding The restriction level parameter limits the cut of the histogram of the sub-block, and then redistributes the number of cut pixels into the gray level of the histogram of the sub-block until all the cut pixels are allocated. (5) Histogram equalization processing: a histogram of each sub-block in which the pixel allocation is completed, and the histogram equalization processing is performed individually according to the HE method. (6) Pixel gray-scale reconstruction calculation: Calculate the average value of the brightness of each sub-block according to the previous step, set it as the reference point, and calculate the sub-area by bilinear interpolation (Bilinear Interpolation). Each pixel in the block points its new gray level.

Another type of technique is to use a way to limit or modify the PDF of the original image to obtain a new contrast-enhanced image. Taking Adaptively increasing the value of histogram (AIVHE) as an example: As shown in equations (7) and (8), AIVHE adds a parameter to prevent image over-enhancement. To limit the upper limit of the PDF, in addition to the benchmark parameters Adjustability parameters with adaptability Through adjustment with The user adjusts the parameters to modify the original image PDF to obtain a new PDF, and then uses the HE method to achieve effective image enhancement contrast. The characteristics of the AIVHE method are as follows: 1. AIVHE utilizes preset limit parameters Avoid the output will have an enhanced visual effect; 2. Adjust the parameters through adaptability Adapt it to contrast enhancement on a variety of different images; 3. through Adjustment can determine the scope of the contrast expansion; 4. through adjustment To make the contrast-enhanced image closer to the original image brightness. Similarly, methods for achieving image contrast enhancement by modifying PDF include BUBO, AMHE, and WTHE. Among them, WTHE has an action of increasing the brightness of the original image to increase the brightness of the original image, so that when the dynamic image is played Does not produce unnatural display results. [Number 7] [Number 8]

As mentioned above, various image contrast enhancement techniques are currently focused on redistributing the entire image by the grayscale value of each pixel to achieve contrast enhancement; however, special image features such as medical contrast images are considered. The common feature of such images is that the grayscale values of the pixels are obviously out of balance, such as a large part of the entire medical image is close to the dark block without brightness (non-medical diagnostic area), and the desired observation The focus area is also the same body tissue and its pixels are scattered in a small gray scale. The dark picture of the camera that is not recorded by the obvious light source at night makes the pixel gray level of the whole image fall to the gray level near the non-brightness part of the whole image. A few bright pixels are often caused by Noise generated by insufficient resolution or line interference. Since such images are distributed by histograms, it is possible that all grayscales have a pixel distribution, but the pixels that really need to be observed are indeed closely concentrated in the grayscale of a certain segment, so if they are enhanced with the traditional contrast. Technical processing, that is, because these technologies are designed with all the pixels of the whole picture in mind, the final enhancement effect is very limited and the content distortion is easily caused.

As mentioned above, the existing image contrast enhancement techniques focus on redistributing the entire image according to the grayscale value of each pixel. When the luminance grayscale value of the pixel of the image is obviously unbalanced, the brightness is locally enhanced. The gray-scale value, if the existing contrast enhancement technology is processed, the effect of the enhancement is limited, and the whole image content is distorted. Therefore, the main purpose of the present invention is to provide a method and apparatus for contrast enhancement of the local image of the histogram. By selecting the brightness range, the gray-scale space of the non-emphasized brightness area is compressed, so that the pixels of the non-emphasized brightness area are squeezed to a small range, and the pixels of the key brightness area can obtain a larger brightness gray level value equalization space, and then Through the specific HS or HE technology, the grayscale value range of the image key block is expanded to obtain more obvious contrast enhancement and image clearness effect, and the entire image is to be locally enhanced to increase its brightness grayscale value. The problem that the luminance grayscale value of the pixels of other areas of the image is unbalanced causes distortion of the entire image content.

The main technical means for achieving the foregoing purpose is to enable the above-mentioned histogram partial image contrast enhancement device, comprising: an input image conversion module for inputting an image and converting it into a YCbCr signal; an area brightness enhancement module The image enhancement module is electrically connected to the input image conversion module, and the brightness enhancement module group converts and generates an adjusted YCbCr signal according to a plurality of control signals; and an output image conversion module electrically connected to the area brightness enhancement module The input image conversion module outputs a YCbCr signal adjusted by the area brightness enhancement module; an image brightness input interface electrically connected to the area brightness enhancement module, wherein the image brightness input interface is used to select more than one Enhanced grayscale range (H, L), or input more than one pixel coordinate (X, Y) and its grayscale range threshold (Z); a standard area brightness analysis unit, which is electrically connected to the image bright input interface The coordinate area brightness analysis unit generates a pixel coordinate (X, Y) and a gray level range threshold (Z) according to the image bright field input interface. a range value (Hnew, Lnew); a bright field generator, which is electrically connected to the input image conversion module, the area brightness enhancement module, the coordinate area brightness analysis unit, and the image bright field input interface, respectively, the bright field generator is The brightness (Y) in the YCbCr signal establishes the probability density function of the image signal, the gray scale range (H, L) generated by the brightness input unit of the image or the brightness analysis unit of the coordinate area, according to the expandable range (Hnew , Lnew) provides a regional brightness enhancement module for contrast enhancement processing.

The main technical means adopted to achieve the foregoing objectives is to make the above-mentioned histogram partial image contrast enhancement method, including: obtaining an image; counting the histogram distribution of the image, and distinguishing each pixel of the image into grayscale values. Regional and non-key areas; compress the non-key areas of the histogram distribution to obtain sufficient gray-scale space to provide the contrast enhancement of the key areas for specific gray-scale areas; Improve the brightness difference and enhance the contrast of its regional image to improve the detail and clarity of the image.

The histogram partial image contrast enhancement device composed of the foregoing components performs contrast enhancement processing on a specific grayscale region instead of enhancing the contrast of the entire image, so first find out the luminance grayscale region of the image, and then compress The histogram blocks of other regions are used to generate sufficient gray-scale space to provide key areas for contrast enhancement, so that the key subjects in the image can achieve better clarity. Since the present invention performs key contrast enhancement processing for image features of specific purposes, the user can select different working methods and different contrast enhancement modes by using the image bright field input interface, and confirm the key gray scale regions and related parameters. Achieve versatile application purposes. To solve the problem that the prior art wants to locally enhance the luminance grayscale value of the entire image, the luminance grayscale value of the pixels in other regions of the image may cause distortion of the entire image content due to the distribution imbalance.

The invention establishes a histogram partial image contrast enhancement device, and for the image recorded by the medical image contrast device and the general monitor in a dark environment, the partial image contrast enhancement is performed through the hardware before the image is output to the display, so that The final output to the display is a clear picture that has been enhanced with emphasis on the local image. Through the device, the user can select two working modes to complete the selection of a specific grayscale region or directly specify the screen block to be enhanced by clicking a specific position of the screen. The technology will adjust the compression/expansion ratio set by the user. Image sharpening processing for grayscale extension of the grayscale region selected by the user: firstly, the histogram distribution of the image is calculated, and then the histogram block of other regions in the distribution is compressed to make enough The gray-scale space provides contrast enhancement of the key areas, and the gray level of the area is enlarged, that is, the brightness difference becomes larger, and the contrast of the area image can be immediately enhanced, so that the details are better and accurately identified to achieve better clarity. Effect. Since the device is directly fabricated by hardware electronic circuits and logic circuits, it has the advantages of high processing speed, and is suitable for the user to adjust parameters or switch operation modes while observing the image effect.

A preferred embodiment of the present invention, as shown in FIG. 10, includes an input image conversion module 10, an area brightness enhancement module 20, an output image conversion module 30, an image brightness input interface 40, and a The coordinate area brightness analyzing unit 50 and a bright field generator 60 are configured to input different image formats (NTSC, PAL, RGB or DVI) and convert them into a YCbCr signal (YCbCr). The area brightness enhancement module 20 is electrically connected to the input image conversion module 10, and the area brightness enhancement module 20 group is provided with a plurality of control signals (threshold value (H), threshold value (L), threshold value (Hnew). a threshold value (Lnew) and a grayscale range threshold (Z), which convert the YCbCr signal according to the control signals and generate an adjusted YCbCr signal (NewYCbCr). The output image conversion module 30 is electrically connected to the area brightness enhancement module 20, and the output image conversion module 30 outputs the YCbCr signal adjusted by the area brightness enhancement module 20. The image brightness input interface 40 is electrically connected to the area brightness enhancement module 20, and the image brightness input interface 40 has a plurality of adjustment parameters (threshold value (H), threshold value (L), pixel coordinates (X, Y). And the gray scale range threshold (Z)), the image bright field input interface 40 is used to select the gray scale range threshold value (H, L) to be enhanced, input one or more pixel coordinates (X, Y) and its gray scale Range threshold (Z); In the preferred embodiment, the gray scale range threshold (H, L) is adjusted by a knob or a switch. The coordinate area brightness analyzing unit 50 is electrically connected to the image bright field input interface 40. The coordinate area brightness analyzing unit 50 generates a pixel coordinate (X, Y) and a gray level range threshold (Z) according to the image bright field input interface 40. Extensible range values (Hnew, Lnew). The bright field generator 60 is electrically connected to the input image conversion module 10, the area brightness enhancement module 20, the image bright field input interface 40, and the coordinate area brightness analysis unit 50, respectively. The light field generator 60 converts the mode according to the input image. The brightness (Y) in the YCbCr signal of the group 10 establishes the probability density function (PDF) of the image signal, and the threshold value of the gray scale range generated by the image bright field input interface 40 or the coordinate area brightness analyzing unit 50 (H, L) The area brightness enhancement module 20 is provided for contrast enhancement processing according to the expandable range (Hnew, Lnew).

The histogram partial image contrast enhancement device of the present invention firstly decodes the analog signal of the medical imaging device or the photographic monitor, first decodes the analog signal such as NTSC into a digital signal, and then performs deinterlacing and other digital signal processing for the video standard, and the signal is processed. The brightness and color difference format (YCbCr) is arranged, wherein the brightness is Y, the blue color difference is Cb, and the red color difference is Cr. The area brightness enhancement module 20 processes the brightness Y, and the processed Y is combined with Cb and Cr. The adjusted YCbCr signal (NewYCbCr) is converted into an NTSC, PAL, DVI or RGB output type through the output image conversion module 30, and then sent to the corresponding display device (screen).

The user performs the following operations through the image bright field input interface 40: 1. Select the mode of use and the compression ratio (Z value) through a knob or a switch. 2. If the dynamic adjustment mode is selected, the brightness key observation area (L, H value) can be adjusted with the knob according to the display result. 3. If you select the specified object mode, you can directly specify one point (X, Y) of the screen image through the input device such as touch screen or mouse, so that the specified object will produce contrast enhancement effect on the screen.

The input image conversion module 10 of the histogram partial image contrast enhancement device is provided with a frame buffer (not shown) for storing the Y value matrix of the image for the coordinate region brightness analysis unit 50 and the bright field. The generator 60 specifies the L and H value analysis of the object and generates the stretch range (Lnew, Hnew), and finally the region brightness enhancement module 20 performs the final image histogram distribution according to the L, H, Lnew, Hnew, and Z values. Stretch/compression.

In the generation of Lnew and Hnew, the gray-scale space of the non-emphasized brightness range of the non-emphasized area is compressed, so that the pixels of the non-key area are squeezed to a small range, and the key brightness range of a large number of pixel sets can be made larger. The luminance gray scale value equalization space is further stretched by the HS technique to make the gray scale value of the image which is originally blurred due to the excessive closeness of the gray scale value; the histogram of FIG. 11 is For example, the gray-scale space of Le to L and H to He in the figure is compressed by 1/Z times and placed in a new position. The new position is the space from 0 to Lnew and Hnew to 255 shown in Fig. 12, in which Le And He represents the lowest gray scale value and the highest gray scale value of the entire original image, and the area smaller than Le and greater than He causes waste of gray level due to non-use, so the space was wasted in the present invention. Move into the stretchable area together. According to the characteristics of the gray scale distribution of medical images, the part that the user (doctor) wants to diagnose is usually concentrated between the H and L of the histogram, so the method sets the area of the gray scale between H and L to be contrast enhanced. The area, the area smaller than L and larger than H is an unimportant compressible area, and the compressible area is compressed as shown in FIG. 12, and the gray-scale space which is originally used less than Le and greater than He is used. The range between Hnew and Lnew is the vacated gray class space.

The stretching/compression technique for the specific region of the histogram distribution is as follows: by determining which region of the histogram the pixel (K) to be processed is located, the corresponding method is selected, and the grayscale portion smaller than the L value is larger than The gray-scale part of the H value, the two parts will perform the space folding compression action according to the input folding rate Z. In the middle of the part larger than the L value to less than the H value, the histogram equalizing the stretching action is performed. The values of equations (9) and Lnew equations (10) of Hnew are used. [Number 9] [Number 10] In formula (9), the maximum grayscale value of the Maxgray table image, K represents a certain pixel in the graph. If a graph is 512×512, then there are 262144 K, and pixel(K) represents K pixel. The gray level, NewPixel (K) represents the new gray level of K. After receiving pixel(K), if the value is determined to be in the L<pixel(K)<H interval of the histogram, the histogram of the interval should be stretched and extended to Lnew to Hnew to achieve the image contrast enhancement effect. Equation (11), Pixel(K)-L represents the gray level difference between this gray scale and L, and multiplied by the stretch ratio of the old range and the new range. And take the smallest integer, you can calculate the gray level difference between the new range and Lnew, and finally add Lnew to get K's new gray class Pixel (K). [Number 11]

Figure 13 is a flow chart of the regional gray-scale HS expansion enhancement algorithm. The grayscale value range to be processed in this flow is assumed to be [0..255], so Maxgray is 255, but the method can also be applied to other differences. Grayscale value range. This process first reads the graph to be processed and puts it into a matrix, and inputs H, L, and Z values: where H and L are to be enhanced by the gray scale range, so that the gray scale falls in the [L:H] range of pixels. By this method, the effect of the extension enhancement is produced, and the Z value is the magnification of the portion of the non-emphasis range (less than L or greater than H) that is compressed toward the two poles in the gray scale distribution of the original image. Then analyze the histogram of the picture and find the minimum brightness value Le and the maximum brightness value He of the picture, then calculate the number of all pixels in the picture into the allelement to be used, k in the flow chart represents a certain pixel, and then use for The loop returns the pixel value pixel(k) of all the pixels of the picture into the formula calculation (k from 1 to allelement), and judges the range of the pixel value Pixel(k) to calculate the new grayscale value of the point in the corresponding processing manner. . k=k+1 means that the calculation of this pixel is completed, and one pixel is replaced. All the pixels are calculated to complete their respective newpixel(k), and the new image is stored in another matrix, that is, the contrast enhanced image is completed.

Each pixel of the image is divided into a key area and a non-key area at both ends according to the gray scale value. The user can select the gray level of the specific position of the specified screen as the reference gray level through the operation mode selection, and use the queue to store the already used Queue. Determining the coordinates of the same block/object as the reference point, comparing whether adjacent pixels belong to the same block/object, and storing the adjacent coordinates of the same block/object in the queue, thus repeating Processing until all pixel coordinates belonging to the same block/object are determined. If comparing adjacent pixels to the same block/object, first determine whether the difference between the grayscale value of the adjacent pixel and the adjacent reference point grayscale value is less than a luminance difference threshold, and if not, the gray of the pixel The order value is averaged with the grayscale values of the surrounding pixels to adjust the grayscale value according to the environmental relationship. If the difference between the grayscale value and the adjacent reference point grayscale value is less than the luminance difference threshold, it is also belonged to the same block. / Object, and adjust the key grayscale range according to its original brightness value. The adjustment focus gray scale range refers to selecting the minimum gray scale value as the L parameter according to the determined pixel in the block/object, and the maximum gray scale value is the H parameter.

14A and FIG. 14B are flow charts for performing area brightness analysis based on coordinates input by a user to generate a brightness range L, H of a specified area/object, and the function is to utilize a point coordinate (x, input by the user). y) Be the starting point, use Branch-&-Bound logic to expand to the square to find the brightness range value (L, H) of the selected area/object, and the image gray to be processed in this process The range of magnitudes is also assumed to be [0..255], but in the future this method can also be applied to other different grayscale value domains. The g(x,y) function is shown in equation (12). The a and b parameters in the formula are the fuzzy interval reference range set by the system, and the w[] matrix is the weighted specific gravity value of the interval range set by the system. This equation is a fuzzy processing formula for the neighboring block (x, y) to make the judgment point (x, y) and the neighboring point (x+i, y+j) belong to a block/object. It can make the judgment of adaptability according to the difference of brightness around, and it produces the L and H values which are more realistic. [Number 12]

For the conversion effect of the image obtained by the present invention applied to the medical imaging apparatus, please refer to the attachments 1 to 3. Since the images produced by the medical imaging apparatus are mostly gray scale diagrams, the images produced by the medical imaging apparatus are often limited due to the performance of the instrument. Not clear enough, although the details are taken, but often the brightness (gray scale) is too close to be visually illegible. Through the histogram partial image contrast enhancement device, it is obvious that the converted image is clear and easy to identify.

The image of Annex 1 is taken by brain tomography (CT). It can be seen from the circled area below that the contrast-enhanced image of the gray matter of the brain (the right circle is selected) and the white matter (the upper circle) are more distinct. The boundary of the tumor (left side circle) and the tumor content are clear. The diagnosis of this picture is to determine the tumor boundary and to examine whether there are calcification, fibrosis or hyperplasia in the tumor, and the tumor boundary and content after contrast enhancement. The matter is more distinct and can help the lesion to judge.

The image of Annex 2 is also taken by brain tomography (CT). The main symptom of this picture is suspected bleeding. The right temporal lobe of the above picture is suspected to have bleeding. After contrast enhancement, the lower right picture can be clearly seen. The shape of the leaf and the left temporal lobe are significantly different (the right circle is selected), and the bleeding can be determined.

The image of Annex 3 is the human lumbar vertebrae image of magnetic resonance imaging (MRI). The above image cannot clearly see whether there is any foreign body oppression in the upper dura mater. The image processing using PACS will make the brighter part of the original more bright, while the middle is hard. The upper surface of the meninges can not clearly affect the interpretation. After adjustment, the image of the lower image can clearly see the upper dura mater (the central circle) and the nerve root (arrow).

10‧‧‧Input image conversion module

20‧‧‧Regional brightness enhancement module
30‧‧‧Output image conversion module
40‧‧‧Image Bright Field Input Interface
50‧‧‧Coordinate area brightness analysis unit
60‧‧‧ Bright field generator

Figure 1 is a prior art original image. 2 is a gray scale histogram corresponding to the image content of FIG. 1. Figure 3 is an original histogram of another conventional image. Figure 4 is a histogram of subtracting the minimum value from Figure 3. Figure 5 is a histogram of the histogram expanded by the histogram extension contrast enhancement (HS) extension of Figure 4. Figure 6 is a histogram calculated by the probability density function (PDF). Figure 7 is a statistical graph calculated by the cumulative density function (CDF). FIG. 8 is a schematic diagram of global full-scale equalization expansion between 0 and L-1 by histogram equalization contrast enhancement (HE). FIG. 9 is a schematic diagram of dividing into two sub-histograms by BBHE and separately equalizing processing. Figure 10 is a block diagram of a circuit in accordance with a preferred embodiment of the present invention. Figure 11 is an original histogram of a preferred embodiment of the present invention. Figure 12 is an expanded histogram of a preferred embodiment of the present invention. Figure 13 is a flow chart of a region gray scale extension enhancement method in accordance with a preferred embodiment of the present invention. 14A and 14B are flowcharts showing a method of brightness analysis of a coordinate area in accordance with a preferred embodiment of the present invention. Annex 1 is a comparison chart before and after image contrast enhancement of brain tomography (CT) images (1). Attachment 2 is a comparison chart before and after image contrast enhancement of brain tomography (CT) images (2). Annex 3 is a comparison chart before and after contrast enhancement of magnetic resonance imaging (MRI).

10‧‧‧Input image conversion module

20‧‧‧Regional brightness enhancement module

30‧‧‧Output image conversion module

40‧‧‧Image Bright Field Input Interface

50‧‧‧Coordinate area brightness analysis unit

60‧‧‧ Bright field generator

Claims (6)

A histogram partial image contrast enhancement method includes: obtaining an image; and calculating a histogram distribution of the image, and classifying each pixel of the image into a key area and a non-key area according to a gray scale value; distributing the histogram The non-key area is compressed, and sufficient gray space is obtained to provide the contrast enhancement of the key area for the specific gray area; and the gray level of the area is enlarged to increase the brightness difference and enhance the regional image contrast. To improve the detail and clarity of the image. The histogram local image contrast enhancement method according to claim 1 is to compress the non-key areas at both ends of the histogram by using a formula versus , so that the key areas extend to both ends. The histogram partial image contrast enhancement method according to claim 2 is to compress the gray scale space of the non-emphasized brightness range of the non-key area, so that the pixels of the non-key area are squeezed to a small range, and the key brightness range of the plurality of pixel sets is concentrated. Obtain a larger brightness grayscale value equalization space, and use HS technology to do the grayscale value stretching, through the formula Convert to the grayscale of the larger segment to expand the grayscale value range. The histogram partial image contrast enhancement method according to any one of claims 1 to 3, wherein the pixels of the image are divided into the focus area and the non-key areas at both ends according to the gray scale value, and the user selects the operation mode. Using the gray level of the specific position of the specified picture as the reference gray level, the Queue stores the coordinates of the same block/object as the reference point, and compares whether the adjacent pixels belong to the same block/object one by one, and Adjacent coordinates satisfying the same block/object are then stored in the queue and repeated until all pixel coordinates belonging to the same block/object are determined. The method for contrast enhancement of a histogram partial image according to claim 4 is to compare whether adjacent pixels belong to the same block/object, and first determine whether the difference between the grayscale value of the adjacent pixel and the gray value of the adjacent reference point is Less than a luminance difference threshold, if not satisfied, the grayscale value of the pixel is averaged with the grayscale value of the surrounding pixel to adjust the grayscale value according to the environmental relationship, if the grayscale value and the adjacent reference point grayscale If the difference between the values is smaller than the luminance difference threshold, it is also subject to the same block/object, and the key gray scale range is adjusted according to the original brightness value, and the adjusted key gray scale range refers to the pixel in the determined block/object. Select the minimum grayscale value as the L parameter and the maximum grayscale value as the H parameter. A histogram partial image contrast enhancement device includes: an input image conversion module for inputting an image and converting into a YCbCr signal; and an area brightness enhancement module electrically connected to the input image conversion module, The area brightness enhancement module group converts and generates an adjusted YCbCr signal according to a plurality of control signals; an output image conversion module is electrically connected to the area brightness enhancement module, and the input image conversion module outputs the area brightness The enhanced YCbCr signal of the enhanced module; an image bright field input interface electrically connected to the regional brightness enhancement module, wherein the image bright field input interface is used to select more than one gray scale range (H, L) to be enhanced, Or input more than one pixel coordinate (X, Y) and its grayscale range threshold (Z); a standard area brightness analysis unit, which is electrically connected to the image bright field input interface, the coordinate area brightness analysis unit is input according to the image brightness field The pixel coordinates (X, Y) of the interface and the grayscale range threshold (Z) produce an extendable range value (Hnew, Lnew); a bright field generator, The two are respectively electrically connected to the input image conversion module, the area brightness enhancement module, the coordinate area brightness analysis unit and the image bright field input interface, and the bright field generator establishes the image signal according to the brightness (Y) in the YCbCr signal. The probability density function, the gray scale range (H, L) generated by the image bright field input interface or the coordinate area brightness analysis unit, and the area brightness enhancement module is provided for contrast enhancement processing according to the expandable range (Hnew, Lnew).