TWI220847B - Fast compression and decompression method for animation - Google Patents

Abstract

The present invention provides a fast compression and decompression method for animation, wherein the compression process includes: converting the pictures in the animation file as a low color level graphic; shielding the same portion therein and clipping to keep the effective portion; generating the color palette based on the features of output equipment; and employing the dictionary compression based on the color levels of effective portion to obtain the animation compression file. The decompression process includes: performing decompression based on the compression dictionary to obtain the decompressed picture; composing a complete picture based on the shielding manner; confirming with the algorithm for recovering to the original color level based on the color palette; conducting the color conversion on the entire picture based on the algorithm to obtain the current picture; and, playing the current picture on the output equipment by pasting the picture.

Description

1220847 V. Description of the Invention (1) [Technical Field of the Invention] The present invention is a graphic compression and decompression method, particularly a dynamic daytime fast compression and decompression method with a high compression ratio. [Previous technology] At present, various portable electronic devices are widely used in everyone's life, such as mobile phones, PDAs, MMVs, handheld computers, and so on. With the continuous development of multimedia technology, due to its various sound and light effects

It should be expected that playing all kinds of multimedia information on the above-mentioned portable electronic devices is a common expectation of users. Moving day files are generally large. In order to be suitable for playing on portable electronic devices, they usually need to be compressed. The existing dynamic daytime compression methods are divided into two categories: 1. Each frame is compressed separately, and then decompressed frame by frame during playback and displayed. This type of method is simple to implement, but the compression rate is often not ideal. 2. Consider the dynamic content as a whole, separate the color signal from the luminance signal and color signal, and then perform lossy compression, such as MPEG. The advantages are The compression rate is high, but the amount of calculation is large. Cannot play smoothly on devices using low speed chips. Most of these compression algorithms are based on a 24-bit RGB color gamut space. For the current large number of handheld devices, the LCD has a color depth of only 1 2b t, which causes a double waste of space and time in data transmission and decompression operations. However, in the various existing portable electronic devices, the hardware conditions provided vary widely. When playing multimedia information on electronic devices, especially animation files, the compressed moving files need to be decompressed.

Page 5 1220847 V. Description of the invention (2) The hardware of electronic equipment, especially the processor and storage equipment, has certain requirements. If you want to play the dynamic day on some low-profile electronic equipment, the playback effect will often be Not very ideal. At this time, a dynamic compression and decompression method with high compression ratio is needed to solve the above problems. [Summary of the Invention] In view of this, the present invention proposes a dynamic decompressing and decompressing method in order to solve the above problems. The purpose of the present invention is to solve the low compression rate and cause space in data transmission and decompression operations in the conventional techniques. Disadvantages, such as the double waste of time, provide considerable dynamic performance and low-cost storage solutions for color models with low hardware configuration. The invention proposes a method for fast compression and decompression of moving day. The compression process includes the following steps: converting all the graphics in the moving day file into low-level graphics; shielding the same part of the low-level graphics; and blocking the graphics Cut it to retain its effective part; generate the corresponding color palette according to the characteristics of the output device; and obtain the moving day compression archive through dictionary compression according to the color level of the effective part. The decompression step includes the following steps: decompress the moving daytime compressed file according to the compression dictionary type to obtain the decompressed graphic; synthesize the decompressed graphic into a complete graphic according to the masking method; and determine to restore the complete graphic to The algorithm of the initial color gradation; according to the algorithm, the complete graphic is converted to obtain the current graphic; the current graphic is played on the output device through the map. Compared with the existing methods, the compression and decompression method provided by the present invention uses a variety of technologies and optimization methods to combine high-speed and dynamic-day fast solutions. 1 1220847 V. Description of the invention (3) Compression / decompression / display scheme High dynamic compression ratio, fast decompression display performance, and make full use of the host's wide color gamut range. This solves the problems of low compression efficiency and waste of space and time in the decompression operation of the traditional method, and makes the color machine playing on the low hardware configuration smoother. The details and methods of the present invention are described below with reference to the drawings. [Embodiment] The present invention will be described in detail below with reference to the drawings: Please refer to "Figure 1", which is a general flowchart of the rapid compression method of animation provided by the present invention, as described below: All graphics in the moving day file Convert to low-level graphics (step 110): mask the same part of the low-level graphics (step 1 20): cut the masked graphic to retain its effective part (step 130): according to the characteristics of the output device Generate a corresponding color palette (step 140): obtain an animated compressed archive through dictionary compression based on the color gradation of the effective portion (step 150). The moving day file contains a series of similar graphics. However, the graphics that make up the moving day are often pictures in storage formats such as true color, such as 2 4b it color or higher graphics. The file size is too large and it is not suitable for configuration. Play in a portable electronic device that is relatively low compared to a computer. Therefore, the data of all graphics in the moving day file must be simplified first, that is, the original graphics in the moving day file are compressed into low-level graphics of 256-color or 16-color graphics to reduce the data volume of the graphics. And data complexity. Against a full color (such as

Page 7 1220847 V. Description of the invention (4) 2 4b color or higher) graphics, we can use dithering or color selection method to convert to 2 5 6 color graphics according to its content, while still maintaining the original graphics Look and feel. Common dithering methods include two-line dithering and Freud-Sternberg dithering; and more commonly used methods include spherical shell color selection. Methods such as dithering or color selection make use of the characteristics of human eyes that are sensitive to brightness information but not sensitive to color information, and replace thousands of colors with a small number of colors, and have approximate display effects, making graphic data large To simplify and shrink. Like other lossy compression methods, this scheme will actually discard some graphics information when converting the graphics to 2 56 colors. The commonly used spherical shell color selection method or Freud-Steinberg dithering color are both preferable. For graphics with a small number of colors, you can switch to 16-color graphics to further reduce the amount of data. The compression process of the present invention will be further described below. Please refer to "Figure 2", which is a flowchart of masking the same part of the low-level graphics according to the present invention, which is explained as follows: When the decompressed graphics uses the key frame masking method, select The first frame graphic is a key frame, and other graphics are used as subsequent frames (step 210): the same portion of the subsequent frame as in the first frame is masked (step 2 2 0), and finally proceeds to step 130. When the decompressed graphic uses the sequential frame masking method, each graphic in the moving day file is treated as a separate frame (step 2 30): starting from the second frame, the same place in each duck as the previous i5 frame is performed. Mask (step 2 4 0), and finally go to step 1 3 0. Graphics with color conversion compression can be obtained when the data content is relatively single

Page 8 1220847

5. Description of the invention (5) If each frame is graphic. And even if it is more advanced, its pressure is a graph with no data redundancy, this value still has a certain related order. Therefore, this article compresses the graphics. The diurnal aspect of the moving side only accounts for the entire diurnal content, so the next > compression ratio based on this idea is more south ^ # Expressive cannot be unavoidable # ^ shape, regardless of the compression calculation, it will tend to Nearly a small _ <-set the specific content of color because the animation is up ~ p-shaped can be regarded as-Jinzhen Mask) technology to simplify the treatment of all possible and its front or back ^ that is, the changing part to us It is known that 20% of the previous diurnal surface can be drawn with large color blocks, and the size of the simple image with a small amount of drawing is only an ideal value); The column of the day surface uses every picture to be shaded. (Shadow for the day, each day surface is the same. For different parts, less than 20% of the surface. It is assumed that the day surface only needs to know the different ones before compression. First

Some graphics are compared for content, and the same redundant part becomes Mask Shadow, so that Mask Shadow occupies most of the area of the graphic to be compressed, that is to say, most of the area in the shadowed graphics file with redundant content removed Both are the same color, which is the color of Mask Shadow. The amount of graphics data after the graphics compression processed as described above is far less than the result of using only color conversion compression.

The masking technology used in the present invention includes two methods: key masking and sequenced mask masking. Suppose there is a group of animations with a total of ten frames. If the next nine frames are compared with the first frame, the average number of the same primitives is 70%, and when compared with the second frame, the average number of the same pixels may be 80%. All of them need to be compared dozens of times. If the number of frames is large, the number of all comparisons increases with the squared relationship with the number of frames.

1220847 V. Description of the invention (6) Among them, the principle of the key frame masking method is as follows: the first frame is used as the key to supplement the 'blessing'. The subsequent blessing is compared with this frame to generate M ask Shadow 5 in the process. Set a Gain value. When the area of a certain M ask Shadow is smaller than this Ga in value, it means that the content of the moving day has changed greatly compared with the previous frame, that is, this frame is also regarded as a key frame without masking. . After that, the day after day is compared with the previous key frame, then the one with a large masking area is taken, and the key 祯 ID used by it is recorded in this mask. The advantage of this method is that each content is relatively independent and can be solved separately. The principle of sequential masking is as follows: assuming that each frame has the highest similarity to the previous frame (which is often the case), each frame generates a masking part for the template. The decompression synthesis speed of this method is fast, but each frame is dependent on the previous frame, which is suitable for sequential playback. After the graphics in the moving day file are masked, it needs to be further cut: only the effective part is retained. The present invention accomplishes this task by using the technology of smart cut. After the dynamic day content has been processed by Shad 〇w M ask, its compression size has been greatly reduced compared to before processing, but for a certain size graphic, even if there is no content in it (such as all white), there is still Certain size. The method of smart reduction is introduced here to further reduce the data. The idea of this method is that if a large continuous area is covered by Mask Shadow, then this area is simply discarded, and only its position is recorded. After this trimming, the area of the masked image is also greatly reduced, and its size after compression is smaller. Sometimes, only a few hundred, dozens, or even zero bytes of frames with high similarity are left. After the cutting process is completed, the color processing stage is entered. The present invention is

_ 醒 _ Page 10 1220847 V. Description of the Invention (7) It is performed by using the color wheel adjustment technology. Because the processed 2 5 6/16 color graphics 'When playing this graphic, the output device may be 4 6 colors, so the corresponding redundant synthesis palette needs to be generated according to the color depth of the output device' to retain this Part of the color scale is mapped to the palette. In the present invention, a process of adjusting a 2 4 bi ΐ true color to a 4 0 9 6 color palette is added here, and a 16-color composite palette and a 256-color parity redundant color palette are generated for 16-color / 256-color graphics, respectively. board. Because not all graphics have 256 colors, for graphics that are not 256 colors, only the palette items of the colors actually used are retained, and unused items are deleted. 2 5 6-color parity redundant palettes are trimmed to take up minimal space. After the color wheel adjustment process is completed, it enters the last step of the compression stage ::. The color level of the effective part retained in the second image is selected. An appropriate dictionary is selected and 2 is compressed into a dynamic day compression file based on the color level storage format. . Since the original figure in the moving day slot case has been converted into a 16 // 2 5 6 pair union, 隹 / order figure in the previous two ^, at this time, it can be based on a dictionary pair based on 16 6 2 5 6 colors. It is compressed. Compressibility i ϊ: ϊ The compression of the canonical model is fast compared to other algorithms, such as LZH and so on. ,:Once. Among them, LZ77, LZSS, LZW, LZA can be used and the degree is different. The differences in the ^, t, and t methods make it sensitive to different types of data. In shrinking the image, the compression ratio of the shape is also not good. The above algorithm has a close relationship. 71 ^ π have a commonality, that is, the larger the product of data complexity and compression ratio, the higher the compression ratio is when the content of a few colors in the graph occupies a larger area. More than ten pairs of animals &% points' This scheme uses a memory cell 1220847 based on 256 colors and 16 colors. V. Description of invention (8) The compression is performed for the following reasons. 4 0 9 6 colors are 12b it In the format, due to the interlaced storage, the primitives are broken, the data complexity is increased due to the interlaced combination, and the compression rate is not high after compression. When using 2 56 colors, one byte per pixel, when there is a large color block, the data is repetitively repeated, and the compression rate is high. For 16 colors, there are only 256 data combinations possible for its 16 colors, and the amount of source data is small, so it can obtain less data complexity and good compression ratio. The present invention is not limited to any specific compression algorithm, but as long as it is one of the above methods or other data dictionary-based methods, this solution can be used to achieve a high compression effect. At this point, the compression process of the graphics in the moving day file is all completed. The decompression process is described below. Please refer to "Figure 3", which is a flow chart of the rapid decompression method of moving day and time according to the present invention, which is described as follows: Decompress the animation compression file according to the compression dictionary type to obtain the decompression graphic (step 3 1 0); Decompress the masking method of the graphic to synthesize it into a complete graphic (step 3 2 0); according to the color palette of the output device, determine the algorithm to restore the complete graphic to the original color level (step 3 3 0); according to the algorithm Perform color conversion on the complete graphic to obtain the current graphic (step 340); play the current graphic on the output device through the map (step 540). When you need to play compressed compressed daily files on various portable electronic devices, you need to decompress them. First, determine the type of dictionary selected for the compression process, and decompress the compressed archive file based on the type. Then determine the masking method selected during the compression process, that is, the solution

Page 12 1220847 V. Description of the invention (9) Masking method of pressing the graphic 'performs the corresponding subsequent operation according to the masking method. If it is determined that the sequential frame masking method is used in the decompressed graphic, the complete graphic is synthesized by direct decompression synthesis (Direct: Unmask Uncompress). If it is determined that the key masking method is used in the decompressed graphic, then the decombination technology will be used To compose a complete graphic. Please refer to "Figure 4", which is a flowchart of decompressing a dynamic daytime compressed file according to the compression dictionary type of the present invention, which is explained as follows: When the decompressed graphic uses the sequential frame masking method, a complete graphic is synthesized by direct decompression. First, Decompress the key frames in the compressed file (step 4 1 0): then decompress the masked frames in the compressed file (step 4 2 0); according to the position information of the cut out masked portion, the masked frame image The primitives and primitives of key frames are combined into a complete graphic (step 4 3 0), and finally enter step 2 2 0. When the decompressed graphic uses the key masking method, first decompress the key frames in the compressed file (step 4 4 0) •, then decompress the mask frames in the compressed file (step 4 5 0); finally enter Step 3 2 0. If it is determined that the sequential frame masking method is used in the decompressed graphic, the complete graphic is synthesized through Direct Unmask Uncompress (Direct Unmask Uncompress). This process is performed simultaneously with the graphic decompression process. If the solved primitives are occluded, they are discarded, otherwise they are pasted to the corresponding position of the previous frame. This method obtains the current graphic after decompression is completed. It's pretty fast. Suitable for moving day in sequence. If it is determined to use the key frame masking method in the decompressed graphics, then the decompression synthesis technology will be used to synthesize the complete graphics. The decompression process is only

Page 13 1220847 V. Description of the Invention (ίο) Extract the key frames and occlusion frames from the compressed file. Please refer to "Figure 5". This figure is a flowchart of synthesizing the decompressed graphic into a complete graphic when using the key mask masking method of the present invention, and the description is as follows: Discard the image element of the masked part obtained by decompression. ); Combining the decompressed non-occluded primitives with the previous frame to obtain the complete graphic of the current frame (step 5 2 0); finally enter step 3 3 0. When using the key frame masking method, when the graphic is decompressed, it is obvious that an occlusion map depends on the content of the key frame or the previous frame to restore its original content. Indirect synthesis is used for dynamic day compression of key frame method. When one or two occlusion maps are decompressed, the key frames are first extracted (this frame may also be buffered in memory for direct use to speed up), and then the occlusion frames are extracted. Combining the unoccluded primitives on the occluded frame with the key frames into a single frame has restored this frame. Because the graphics are converted to 2 5 6/16 colors during the compression process, which is different from the color depth of the output device, you need to convert the data to the output device color depth when decompressing. To reduce the amount of calculation, the unmasked part can be converted based on only the masked information. When the compressed graphic is synthesized into a completed graphic, first enter the step of generating a corresponding redundant composite palette based on the color depth of the output device, and formulate the algorithm steps of restoring the color gradation in the complete graphic to that in the original graphic . The present invention uses a fast color transfer (F a s t M a p P a 1 e 11 e) technique to formulate this algorithm. The graphic data obtained through the above decompression are still 16-color or 256-color graphic data. In order to be able to display it finally, it must be converted into 4096 colors. 1220847 V. Description of the invention (11) The general color conversion method maps 16 colors to 4 0 96 colors through the color palette. It is necessary to take 16 color data (4 bit / pi X e 1) out of one byte. , Then find its corresponding color in the palette (1 2 bit / pi X e 1), and then synthesize this 1 2 bit data to the target graphic. This involves several byte splits, resulting in low synthesis and conversion efficiency. In order to achieve a fast color conversion process, a 16-color composite palette is carried in the compressed graphics, avoiding bit operations. For the so-called synthetic palette, for each 1-bit color, each pixel occupies 1.5 bytes of space. Then, the two pixels have a total of 3 bytes. In the future, the two pixels will be regarded as a color operation. The source data One byte is operated at a time, and the target data is operated at 3 bytes at a time. Since any two pixels of sixteen colors are combined into one byte, it may have only 256 types, so a palette of 256 items can be synthesized in advance, so that two pixels use one palette item , So the conversion process becomes: take two pixel data (one byte), find its corresponding color item (2 4b it, three bytes) in the palette, and write the color item to the target The graphics completely avoid bit operations and greatly improve the conversion efficiency. For 2 5 6 colors, the color palette is too large if the color is converted to a synthetic color palette. Here, the parity redundant color palette technology is used. The elements of this method are: each color creates a 4-byte palette, two bytes are odd pixels, and two bytes are even pixels. Used alternately to reduce bit operations to achieve high speed. In order to further increase the speed, for the 16-color conversion, the above two methods are combined to form a redundant composite palette, which has further improved the performance on the CPU. The design idea of the redundant palette is to use 1 6 b i t to place a 1 2 b i t

Page 15 1220847 V. The palette item of the description of the invention (12) can avoid bit splitting when this color item is taken out from the palette, thereby reducing the calculation time. On a 32-bit processor, the CPU accesses the data bus (BUS) — fetching 4 bytes at a time. If you want to take three, you need to add some operations to three. Therefore, the color items of the synthesized palette are also processed redundantly, and the 24-bit color items are placed using 4 bytes, which can further speed up the color conversion operation. After the algorithm is formulated, the synthesized graphics will be displayed after color conversion according to the algorithm. The present invention provides two methods for color conversion, one is a global color conversion method 'and the other is a local color conversion method. The Global Area Map Palette method takes advantage of the fact that the key frame contains image data not found in other frames. Therefore, the global frame color palette method is used for the key frame, and the entire frame image is converted with the redundant synthesis palette. . The Local Area Map Palette method makes use of the dependence of the cut graphic on the previous frame. As can be seen from the foregoing techniques, for the sequenced mask method, each frame of the graphic is a part of the previous frame. change. Therefore, as long as a clever method is used here with the redundant composite palette to change the color of some of the primitives that are changed, you can get correct results. Because the graphics primitives in this part are only a small part of the total graphics area. Local color conversion greatly reduces the data of color conversion operation. As a result, the color transfer speed is significantly increased. Finally, the current graphic is played on the output device through a map. The present invention provides a global map (G 1 〇 b a 1 A r e a BitTransfer) and a local map (Local Area BitTransfer)

Page 16 1220847 V. Description of the invention (13) Two methods. For key frames, use the global mapping method to throw the entire image into the display buffer for playback. The method of local mapping is similar to the method of local color conversion. When we finally display it, we don't have to throw all the graphic data to the screen. We only need to throw the changed part to the screen to get the correct result. Because this part of the picture element is only a small part of the total picture area. Local maps greatly reduce the data for map operations. As a result, the mapping speed is also significantly increased. At this point, the process of decompressing and playing the compressed moving day graphic file is all completed. The following uses a preferred example to explain the specific process of shielding and smart reduction in the present invention. Please refer to Figure 6, which is a schematic diagram of the masking and smart reduction mentioned in the present invention, which is explained as follows: The four graphics in the first row are the original graphics contained in the animation file. There are quite a few similarities. The four graphics in the second row are masked graphics. The key frame method is used here to mask the same place in the four original graphics. The first frame is used as a template, and the color of the masked part is assumed to be black. It can be seen that from the second frame, the same parts as those in the first frame are replaced by the masked parts. This greatly simplifies the contents of the materials of the next few towels. In other words, the efficiency of compression will be greatly improved. The four graphics in the third row are smart cuts of the masked graphics

Page 17 1220847 V. Description of the invention (14) Subtracted result. : As shown in the figure, only a few small tiles with different contents from the key frame are left in the drawing after the reduction, and the amount of data is greatly reduced. In summary, the above-mentioned ten key technologies of innovation are not used in combination. The present invention provides a high dynamic day compression ratio (average of less than 1 kbytes per frame per day). Fast playback performance (16 0 * 1 60 animations have exceeded 10 frames per second) and as much color use space as possible (because of the Neshin palette, the artist can choose as many colors as needed, And not necessarily limited to standard swatches). The dynamic day compression and decompression implemented through the above process can achieve a high compression ratio of 10: 1 --- 3 0: 1 for general dynamic day, and the animation playback is faster than other existing compression methods. Play 160 * 160 size color animations <10fps on a low-end processor (Toshiba T 9 0 (18MHZ) platform. The above is only a relatively small implementation of the present invention. The examples are not intended to limit the scope of implementation of the present invention; that is, all equivalent changes and modifications made in accordance with the scope of the patent application for this creation are covered by the scope of the invention patent.

1220847 on page 18 Brief description of the diagram. The first diagram is a general flowchart of a fast-day compression method according to the present invention. The second diagram is the process of masking the same part of the low-level graphics according to the present invention. Figure 3 is a flow chart of the fast daytime decompression method provided by the present invention; Figure 4 is a flow chart of the present invention based on the compression dictionary type to decompress the moving daytime compressed file; Figure 5 is a flow chart of the present invention A flowchart of synthesizing decompressed graphics into a complete graphic when using the key frame masking method; and Figure 6 is a schematic diagram of masking and smart reduction of the present invention. [Illustration of graphical symbols] 110 Convert all the graphics in the moving day file into low-level graphics; 120 Mask the same part of the low-level graphics 1 3 0 Cut the masked graphics and keep the valid part 140 according to The characteristics of the output device are used to generate the corresponding color palette. 150 According to the color gradation of the effective part, the moving day compression file is obtained through dictionary compression. 210 The first frame graphic in the moving day file is selected as the key frame, and other graphics are used as subsequent frames. 2 2 0 pairs In subsequent frames, the same parts as in the first frame are masked. 2 3 0 Use each image in the moving day file as a separate frame 2 4 0 Starting from the second frame, mask the same place in each frame as the previous frame 310 Decompress the moving day compressed file according to the compression dictionary type to obtain Unzip

Page 19 1220847 Schematic illustration of the graphic 3 2 0 The decompressed graphic is synthesized into a complete graphic according to the masking method of the decompressed graphic 3 3 0 Based on the palette of the output device, the algorithm for restoring the complete graphic to its original color level is determined. 34 0 Convert the complete graphic according to the algorithm to obtain the current graphic 3 5 0 Play the current graphic on the output device through a map, 410 decompress the key frames in the compressed file 4 2 0 Decompress the occluded frame 4 3 0 Combine the occluded frame primitives and key frame primitives into a complete graphic according to the position information of the cropped occluded part 44 0 Decompress the key frames in the compressed file 4 5 0 Decompress the masked frames in the compressed file. 510 Discard the masked parts of the decompressed image. 5 2 0 Combine the decompressed non-obscured graphics with the previous frame to get the complete graphic of the current frame.

Claims (1)

1220847 VI. Application for Patent Scope 1. A moving day will be based on the compressed file 2. According to the compressed file 2, such as the patent application will be the day of the graphic 3, such as the patent application will be the day of the graphic 4. If the party who masked the key frame in the patent application selects it as the subsequent pairing 5. If the patent application masks the key frame in the order, please shake the amount of data in the range file Please select the amount of data in the range file Please range low level Use the low-level mode to perform a fast compression method in a continuous low-level mode. The method includes the following steps: All graphics in the day file are converted to low-level graphics; the same portion of the low-level graphics; Cut the over-shaded graphic to retain its effective part; generate the corresponding color palette based on the characteristics of the output device; and the color level of the effective part, use dictionary compression to obtain the moving day fast compression method described in item 1, All graphics are converted to low-level graphics. Color processing is used to convert low-level graphics to low-level graphics, reducing the speed of the animation described in item 1 of 0. Method, all of its graphics are converted to low-level graphics by color processing to convert them to low-level graphics, which reduces the dynamic fast compression method described in item 1 above, and the steps of the same part of the graphics are passed Turning off masking includes the following steps: The first frame of the case is a key frame, and the other parts of the figure are the same as those in the first frame. The moving day fast compression method described in the first item, the steps of the same part in the figure are through the secondary masking, which further includes the following steps: Page 21 1220847 6. Scope of patent application Each graphic in the moving day file is treated as a separate frame; and starting from the second frame, the same place in each frame as the previous frame is masked. 6. The quick day-to-day compression method as described in item 1 of the scope of patent application, wherein the masked graphics are cut, and the effective part is retained by cutting the blank part and the masked part in each figure. , And record the position of the cropped part. 7. The rapid day-to-day compression method as described in the first item of the scope of patent application, wherein generating a corresponding color palette according to the characteristics of the output device is to generate a corresponding redundant composite color palette according to the color depth of the output device. The gradations of the remaining parts are mapped onto this palette. 8. The fast day-to-day compression method as described in item 1 of the scope of patent application, wherein the animation compression file obtained by dictionary compression based on the color level of the effective portion is selected according to the color level of the effective portion retained by the graphic. Dictionary compresses it into a moving day compressed file based on the color gradation storage format. 9. A dynamic daytime rapid decompression method, the method includes the following steps: decompress the dynamic daytime compressed file according to the compression dictionary type to obtain a decompressed graphic; synthesize the decompressed graphic into a complete graphic according to the masking method; and according to the color palette of the output device 'Determine the algorithm to restore the complete graphic to the initial color level; perform color conversion on the complete graphic according to the algorithm to obtain the current Page 22 1220847 VI. Patent scope graphic; and play the current graphic on the output device through a map. 10. The dynamic daytime rapid decompression method as described in item 9 of the scope of patent application, wherein the step of decompressing the dynamic daytime compression file according to the compression dictionary type, and the step of obtaining the decompression graphic further includes when the decompression graphic uses the sequential frame masking method, by directly The steps of decompression and synthesis to complete the graphics include the following steps: decompress the key frames in the compressed file; decompress the masked frames in the compressed file; and mask the frames according to the position information of the cut out masked parts Primitives and primitives of keyframes are combined into a complete graphic. 1 1. The dynamic daytime rapid decompression method as described in item 9 of the scope of patent applications, wherein the step of decompressing the dynamic daytime compression file according to the type of compression dictionary, and obtaining the decompression graphic further includes when the decompression graphic uses the key masking method, including The following steps: Decompress the key frames in the compressed file; and decompress the masked frames in the compressed file. 1 2. The dynamic daytime rapid decompression method as described in item 9 of the scope of patent application, wherein the step of synthesizing the decompressed graphic into a complete graphic is performed when the decompressed graphic uses a key frame masking method. The step of synthesizing the completed graphic further includes the following steps: discard the primitives of the masked part obtained by decompression; and synthesize the non-occluded primitives obtained by decompression with the previous frame to obtain a complete graphic of the current frame. Page 23 1220847 VI. Scope of patent application 1 3. The dynamic daytime decompression method as described in item 9 of the scope of patent application, where the palette is a corresponding redundant synthesis palette based on the color depth of the output device. 14. The dynamic daytime rapid decompression method as described in item 9 of the scope of patent applications, wherein the complete graphic is color-changed according to the algorithm, and the current graphic is obtained by performing local color conversion on the graphic restored by color gradation. Method to get the current graphic. 15. The dynamic daytime rapid decompression method as described in item 9 of the scope of patent applications, wherein the complete graphic is color-changed according to the algorithm, and the current graphic is obtained by performing global color conversion on the graphic restored by the color gradation. Method to get the current graphic. 16. The fast daytime decompression method as described in item 9 of the scope of patent application, wherein playing the current graphic on the output device through a map is to display the graphic in the daytime on the output device through a global map method. 17. The fast daytime decompression method as described in item 9 of the scope of patent application, wherein playing the current graphic on the output device through a map is to display the graphic in the daytime on the output device through a local map method. Page 24