TW560174B - Method for compressing an image by using characteristic values - Google Patents

Abstract

The present invention relates to a method compressing image data, more particularly to a method for compressing image data by using characteristic values. At first, input image data are executed by using a pre-processing process of the present invention. The pre-processing process is to use a mathematical calculation method, a filter, or a database to differentiate outlines of objects and to classify different objects of the image data. Then different colors, which are in the outlines of the different objects, are acquired by using a color classifying method and the different colors are used to be different characteristic values of the image data. At last, the different characteristic values are saved individually in a memory to finish the method for compressing image data by classifying colors.

Description

560174

5-1 Field of the Invention: The brake system ^ ^ is a method for compressing images, and in particular, it relates to a method of compressing images with a value of %% 1 to reduce the amount of data and increase memory. Images stored in the card The amount of data. 5-2 Background of the Invention: ^ & Today's information users, especially personal computer users, ^ The mode of computer corruption has entered from the more monotonous text mode to "Month of the Moon Media (Multi- Media) era, and multimedia means that the data contains text, video, sound and other content, and in order to make the data appear in front of the audience in a multimedia style, there have been developed a variety of retrievable images. Multimedia device. Because the multimedia data contains more internal a's, the memory capacity required to process the multimedia data should be higher than that of the processing mode, so that the multimedia data can be presented in a smoother mode. For image input devices, scanners and digital cameras are more common devices. Digital cameras use digitized information to store captured images, and use a disk that can record images or the Personal Computer Memory Card International Association (ASSOClatón; PCMCIA). Memory card to store still images. in order to

560174 V. Description of the invention (2) Compression standard for still images taken by digital cameras (JPEG) file format. The printer can also be used via the RS-232 communication port and no negatives are installed because it can send files to the newspaper. There is a gap, especially inferior to traditional cameras. (Liquid Crystal photos, and this is a traditional camera than a computer, you should use the video :: The memory space used by the media, usually (kint: compressed file To store, common are otographic Experts Group; print by or print, and the problem is that the pre-digital phase of the photo is not only digital display in the light source; the image taken is not comparable with the scanner In the middle of running a camera print, the LCD screen of a traditional film camera is used to immediately view the display screen to process the real photo through the image processing steps. The reporter made the e-mail camera clear the display screen. The input port of the digital camera for the digital camera input computer does not need to be traditionally very suitable for the resolution and sufficient place of the direct image archiving machine. Digital cameras usually have LCD), you can also pass through this. If we want (scanner) (Another name for digital cameras is "Digital Still Camera (DSC)", which is used for digital motion photography. The difference between cameras is that most digital cameras currently use charge-coupled device (CCD) or complementary metal-oxide semiconductor (CMOS) technology to capture images.

Page 6 560174 V. Description of the invention (3) The first image A and the first B of Lingzhao are shown in this figure. The first picture A is drawn in the traditional (green) #B (Mue) field. In the first image A, each square represents a color data of the digital denier / derby material starting area, and the number in the block ~ the position on the previous image. In the description of the present invention, the pixel value of the dagger pixel is represented by the symbol P (x, y), and the pixel value of the first pixel in the column of straight: plain shirt image data-figure number). 11 (η is natural and the first B picture is shown to correspond to each, and the parent pixels ^ and 6 data are respectively f subjects. Then the dimension of the data generated by the parent pixel is = 2. "⑴ means, From the first chart A and the chart _B: When it is expected to arrive at the personal computer, it is to go out. And the assets to be transferred are stored as above, when the buffer memory will store the assets stored in it. Generally come Say, the picture 'so if it is not processed, it can be seen on the digital camera's memory card. The traditional scanner will first store all the R, G, and B data of each pixel in the image before it is stored on the digital camera. When the storage limit is reached slowly, the buffer storage is transferred to the memory card in the digital camera = the memory space occupied by the pixel is quite large. The R, G, and B data of each pixel are all cast. The memory card of the digital camera will Will often reach page 7 560174 V. Description of the invention (4) Go to its data store. When the capital is limited, you must transfer a lot of the data stored in it. When the car driver is big, the memory card of the digital camera is washed. This will shorten the use of digital cameras when n is fully loaded, compared with devices and memory cards: t m: and will shorten the needs of the buffer storage and i: odd life. Although the memory card of the digital camera can be used], 疋-easy to cause the user's economy ::: the stored image data, shrinking can add digital If the camera moderately presses the image to the number of images that can be stored in the digital camera ’s memory card. 5 — 3 Purpose and summary of the invention: For the background of the invention described in the above, the traditional data storage methods, "The amount of stored data is too large for the customer. The second method is to set the formula m: the time it takes is too short. The present invention mainly compresses the data first, and the Πτ knife compresses the video data by compressing the image data. In order to reduce the data amount of the image data. The second objective of the present invention is too. The M 1+ method first uses the feature value to compress the number of image data. I,, s to increase the memory card storage image The second method of the present invention is to reduce the number of times that the memory card is fully loaded. The method is to compress the image data by using the feature values. 豕 贝 # 先 仃 compress, yiyi: on the grid ,、 分 1 ,,, 1 The fourth item of the present invention Is the use of eigenvalues to compress image data

560174 V. Description of the invention (5) Method of precompressing image data 'to speed up data transmission. Still another object of the present invention is to compress the image data in advance by using the surname and method :: = to reduce the length of the image data to extend the service life of the 5H card. According to the above-mentioned purpose, the value of this issue is to reduce the amount of data used by the source of the compressed image data. The present invention first inputs the scene material; the steps of the image processing. The previous processing steps are the benefit side: -Front changer, or database for object rotation.- Method, format conversion of different objects. When the input image data: After classifying the image data, the input image data will be simplified: the processing steps of the object Image data, each of which is equal to a number of colors with different contours. Next, use color points: the object of the contour contains different colors contained in the color, and this is'. Take the contour of different objects The color of this color classification is used as the characteristic value of the image. The difference is to rely on the different shades of different colors or the characteristics of the color type to be stored in ―: self; the two members finally differ in the image data. No .: The side of the material ★. This ⑯ distinction; 成: The characteristic value in the image data memory that uses the seam compression image input is representative of the occupation of the data = the method of the present invention can reduce the storage of image memory cards Store Storing image data ::::. The method of the present invention can also increase the number of times that the t memory card is fully loaded. This issue 0 >. Speed of bucket transmission and prolong memory ^ Use life 560174 V. Invention Explanation (6) 5-4 Detailed description of the invention: Next. However, in addition to the detailed embodiments, the scope of benefits shall prevail. Some embodiments of the invention will be described in detail as described in t. The invention can also be widely used in other inventions. The scope of is not limited. It uses the following method to issue a copy of the original memory to convert an image data in a previous format. The signal is processed differently before the signal is judged to be a database before the invention. When the first film is lost And its color contains a color—a kind of profitable material. This step. Previously, the different objects in the database are in accordance with the regulations. One of the main objects in the first process of the image data will simplify the image data. Describe the process of providing a holographic stepper with less image volume, or in-material or data. The input and processing are the second feature material 1 of the first image data inserted and indicate the first processing. The so-called standard process is to use the contours or material passes for image data. The median value is to reduce the storage. The input steps are formatted as contours. The classification of each object is not counted. The shape is pre-processed with a majority. Each non-reduced image is stored in the first use of the judgment converter. The shape of the device is the same as that of the object and format. The steps are different. The method of calculating the image data when the square data is the same as the contour data. It is a sort of number. In the present invention or the outline of the contour, therefore, after the converter, or identification and division, the objects of the contoured objects are included in the object, and then the color classification method is used to retrieve the contours of different objects.

^ 00174 -----, the description of the invention (7) contains more than one of these colors, and uses the different colors as the characteristic values of the image. The difference is to use the knife's method to use the lightness or darkness of the color. The same type of characteristic value, the same color classification. Finally, the non-image data in the second image data is stored in a memory to complete the compression of the memory by the feature value. Eight methods, in which the memory used in the present invention can be a flash-second image-the feature value stored in the memory can represent the input of the input material]; therefore, using the present invention The method can reduce the storage of image data. = 5 The space of the body of memory and the value that can be stored in the memory to replace # 2 bad. By the method of the present invention, the characteristics of the stored image data are used. From the memory, the amount of image data can be reduced. Therefore, when less ^: the image data is transmitted to the computer, the reduction of the amount of data can reduce > and the time of data transmission. Silly times, shown in the figure, this is a flowchart of using the feature value to compress the shadow 1 η / shell material in the present invention. In the present invention, an image data is input first, and then the input image data is subjected to a pre-processing step 丨 丨 ○ knives ~ images = different objects in the shell material. Next, use the color classification method A # | take + [S] # different colors contained in the outline of the piece ’i use this different color as the feature value 120 of the image. Finally, the differences in the image data are stored in a memory 130 to complete the method of compressing the image data by the feature values. The characteristic values stored in the memory can represent the input image data. The following is the use of the method of the present invention to compress image data with eigenvalues.

560174 V. Description of the invention (8) Material-embodiment, but does not limit the scope of the present invention. When the first image material is rotated, the first object of the present invention is Its * th-object contains the image; 枓 cut-the first step of the processing of the first shape of the neodymium neodymium shape, the first shape inside the sentence is the first one. The second object contains the second shape, the second step of the skin:- The color is not the same as the second color, and the first shape s is not the same, and the first-方 square, I ^ are the same, so the previous processing steps are $, 丨 open method, 转换 器 converter, or / is Utilize the first and second objects of the person's first and second assets. When the wheel image data can be processed, the first-image-color second image data is input. The color of the brother makes the color classification method take the first-color in the first object and it becomes the first special steam 2; and the second color in the second object is captured to be stored in the memory. Finally, the first eigenvalue and the second eigenvalue are respectively stored in the memory. I. The memory used in the present invention can be a flash image data. One eigenvalue and the second eigenvalue can represent another input of the first material. This is a mere example of using the method of the present invention to compress the image data with the eigenvalue, but it does not limit the scope of the present invention. Page 560174

Ming Shou :: C :: ϊ If the material contains a number of different objects, the temple, then this post—mostly ^ The first image data is subjected to a pre-processing step, which is colored. Because any plural number = two of this shape, and this shape contains a method of performing two noses, a converter, or the same object in the database. Sue people & & = most

The image data input from the dry field makes τ < the image data input by the old J Tian Shangren H can be simplified into a package; number :; after the second step, input the data. The second image of several different objects is then used to make the most of the colors in the ancient shape of the color classification, so that the number of characteristics is stored in the memory. value. Finally, the memory can be a flash memory. This $ plus ^ is the first image data used in the present invention. In this case, several characteristic values can represent the input. The following is another embodiment of using the material of the present invention, but it is not limited to the scope of the present invention by compressing the image with a special policy value. When the inputted first image data includes an invention, the inputted first image data is first entered into a pot and a potted plant, then the shapes and contours of the human face and the potted plant are greatly pre-processed. Because the steps are to use the method of differentiating, the type conversion is the same as the judgment of the contours of the previously processed pieces, to classify the image data meat tool, or the database to carry objects _, / α face and bonsai. Needle Office 560174 V. Description of the invention (10) $ —The shape of @ 1 官 and the color of eyes, ears, nose and the first shape in the video data. $ z 颜 For the first pot, the flower pot is divided into a flower shape and the plant material has been shaped to include the first color, the third color, and the third image material. The pots and plants have the first shape, six shapes, and five colors. With six shapes. When the first input, the first input, the third shape, the second color, the second image data < the contours of the human face are not the same, the eyebrows, and the nose have the first shape. There are two parts of the outline and shape of the pot material in the ear material. The six colors are the same as those in the second shape, the first color and the sixth color after the first step. In terms of the previous steps, they are different. In the steps of the medium flower pot, the face is thinner. The eyes have two shapes. Mouth color and the fourth processing step Therefore, the five colors of the basin and the image data will be the fourth shape and the third color. Because people are divided into eyes, the first face has a shape. In the image, it can be simplified because it is planted in a thin fifth shape image. The fifth and fourth colors are the next color, the third color is the first color and the first color is the first color. The color of the eye and the mouth are different. The potted plants can be used for color classification, and the fourth, first, and third eigenvalues are eigenvalues and methods for color classification. The lightness and darkness use it to say, the flowerpot method captures the sixth color, the second eigenvalue, the characteristic value, the eigenvalue, and the color types, including the eye, ear, and value. The color, value, first, fourth, and sixth colors of the targeted color are classified as the special color and the fifth and second color of the plant. The color of the second color becomes the data value of lightness and darkness. The color is used as the available color. The first and fourth eyebrows are the fourth and sixth special faces, and the eyebrows have different characteristics. The second and third colors are different. The eyebrows can be used for eyes, nose, and video.

560174 V. Description of invention (11) Levy value. Finally: the first eigenvalue, the third eigenvalue, the fourth eigenvalue, and the fifth eigenvalue in the image data are stored in the memory to complete the use of the special desire value to store the memory used in the present invention. A quick blink ^ value, the second feature value, the third feature value, the fourth feature value, the first feature value and the sixth feature value can represent the first image data input. The following is another embodiment of using the method of the present invention to compress image data with characteristic values, but does not limit the scope of the present invention. When the first step is different, the contours of the image are the first data that the face is subdivided into. The input image will be the input face. The shape of the face will be judged by calculation. The image data can be simplified into the first image shape and contour method of each unlike the data package, the first face shape and contour pattern classification of the first face, and the nose passing most of the same contours »with most data The image data of the phase converter is different in the previous steps, the eyebrows, and the steps are different. The inner object is different from the previous one, because it is the same, or because it is not processed, so it is behind the mouth. In the case of facial expressions, the steps of processing can be performed in the same steps in the previous database. Because of the rational walking objects. For, because each face has a special part. When inputting the second image color of the input first image file. Next, use the color classification method to capture the different colors contained in the wheel of different objects, and use this different color as the feature value of the image.

Page 15 560174 V. Description of the invention (12) Where the color classification is different to distinguish different colors, eyes can be eyes, ears, nose, and mouth values. Each face is composed of a color classification process that contains a number of feature values. Finally, the second device is stored in a memory, and the method 'wherein the present invention is stored in the memory image data separately. The method is for color classification, and the bright and bright colors can be used for a group of special features. Therefore, the image data can be used to complete the color used in the body. For the shadow darkness as a combination of its different characteristics, after passing through the material, the characteristic values of the memory that can be used by the characteristics contained in the material, that is, the color of the characteristics of the lightness and darkness image data, and the multi-valued compression contained in the processed multi-array Or each value of Yan 'and as an array of different types of images with different types of facial features, such as faces and eyebrows, characteristic feature values, and feature data, the first image can be a flash memory and can represent Enter the first aggregated image volume. Step. This is a data object. Image data, its color separation, and similar methods of the above-mentioned pre-processing library of the present invention. When the input material will be available to each type of party, this will not be profitable. Simplify the extraction of the same color for different rounds. Use the color data to provide the input data for the use of the profile. The use of multi-profile objects with different colors as a light and shade method provides a method for compressing the image data to determine In the past, the degree of several round images without inner packages or the steps of the previous method and format classification of image management in order to reduce the amount of material included the same contours with their colors. The characteristic values contained in the profile are based on the image processing data converter or the difference in the material. After the input image data of the input image is different, this color is different in different colors. The color classification is the same as that on page 16 of 560174 V. Description of the Invention (13). Finally, it is stored in a memory to complete the separate storage in the memory, so the space of the memory of the present invention is used. The reduction of the number of sheets of this material can speed up the speed of data transmission. In addition to its practical effects, and its step-by-step improvement, it is already in line. For this reason, we appreciate your trial and error, and you will feel your sense of virtue. The characteristic value image data of the input image data, the memory card stores the useful life of the present invention, has the function according to the law and the text, prays for the early gift of the image into the memory card using the method of the body full and extended as previously The eigenvalues and eigenvalues in the patent inspector's data can be reduced. Different compressed images to be examined in detail, which are not included in the second-longest memory method, are the number of stored images that can be increased. . The design of the card, the pieces, the inspection, and the remaining methods, respectively. The method of storing images occupied by image resources is not only specific and suitable for application. The above description is only a preferred embodiment of the present invention and is intended to illustrate rather than limit the scope of the patent in the present invention. The essence of the invention is hidden. Without departing from these changes should still be = still :: implemented by ^, please define the scope of the patent in the following. 0This ’m fan addiction is caused by

Page 560174 The diagram briefly explains the above-mentioned objects and advantages of the present invention. The following embodiments and diagrams will be used to explain in detail as follows. Among them: The first diagram A shows how to select the digital image data of R, G and B. The first diagram is the schematic diagram of the situation corresponding to the R, G, and B data generated by each pixel; and the second diagram is the flowchart of the method of compressing images using feature values according to the present invention. . The main part of the symbol: 1 The first block 2 The second block 3 The third block 4 The fourth block 5 The fifth block «6 The sixth block R (X, y) image data of red pixels- Pixel value G (x, y) Image data of green pixel-Pixel value B (x, y) Image data of blue pixel-Pixel value X column coordinate y row coordinate

Page 18 560174 Schematic description of the number of pixels that can be accommodated in block k Number of block η Number of pixel values in block

Claims (1)

560174 6. Scope of patent application 1. A method for compressing image data by using feature values, comprising: capturing a first image data, the first image data includes a first object and a second object, wherein the first object Including a first shape and a first color, and the second object includes a second shape and a second color; capturing the first shape, the second shape, the first color, and the first image data in the first image data The second color to convert the first image data into a second image data; capturing the first color into a first feature value and capturing the second color into a second feature value; and storing the first color The characteristic value and the second characteristic value are in a memory. 2. For the method of compressing image data by using feature values in item 1 of the scope of patent application, the above-mentioned memory may be a flash memory. 3. The method of compressing image data using feature values as described in item 1 of the scope of patent application, wherein the first shape and the second shape described above can be calculated by a method. 4. The method of compressing image data using feature values as described in item 1 of the patent application, wherein the first shape and the second shape described above can be retrieved by a format converter. 5. The method of compressing image data using feature values as described in the first item of the patent application, wherein the first shape described above is different from the second shape. Page 20 560174 6. Scope of patent application 6. — A method for compressing image data using feature values, including: capturing a first image data, the first image data contains a plurality of different objects, any one of which is a majority Different objects include a shape and a color; sequentially processing the plurality of different objects to perform an image data compression process, wherein the image compression process includes: capturing the shape and the color to convert the first image data into a Second image data; capturing the color to make it a feature value; and storing a plurality of feature values of the plurality of different objects in a memory. 7. For the method of compressing image data by using feature values in item 6 of the scope of patent application, the above-mentioned memory may be a flash memory. 8. For the method of compressing image data by using the feature value in item 6 of the scope of patent application, the above-mentioned shape can be retrieved by calculation. 9. For the method of compressing image data using feature values as described in item 6 of the scope of the patent application, the shape described above can be retrieved by a format converter. 1 0. A method for compressing image data using feature values, including: capturing a first image data, the first image data includes a human face, wherein the human face includes eyes having a first color and a first shape, A nose with a second color and a second shape, a mouth with a third color and a third shape, Page 21 560174 6. The scope of patent application and ears with the fourth color and the fourth shape; capture the first shape, the second shape, the third shape, the fourth shape, The first color, the second color, the third color, and the fourth color to convert the first image data into a second image data; capturing the first color into a first feature value, capturing the The second color becomes a second feature value, the third color becomes a third feature value, and the fourth color is extracted as a fourth feature value; and the first feature value, the second feature value, the first feature value are stored. The three eigenvalues and the fourth eigenvalue are in a memory. 11. The method for compressing image data by using feature values according to item 10 of the scope of patent application, wherein the above-mentioned memory can be a flash memory. 1 2. The method of compressing image data using feature values according to item 10 of the scope of patent application, wherein the first shape, the second shape, the third shape, and the fourth shape described above can be extracted by calculation. 1 3. The method of compressing image data using feature values according to item 10 of the scope of patent application, wherein the first shape, the second shape, the third shape, and the fourth shape can be retrieved by a format converter. Page 22