490973 經濟部中央標準局員工消費合作社印製 A7 B7五、發明説明（/ ) [發明之技術領域] 本發明係有關一種影像處理系統，明確地說，本發明 是一種有關向量量化處理技術的影像壓縮系統，更進一步 地說明’本發明係一'種利用簡單的灰色關聯運算方法，尤 適用於即時的影像壓縮應用，在人眼可接受的失真範圍 內，可大幅節省影像儲存裝置的記億空間，並能應用在各 種多媒體裝置。 [發明背景] 隨著電腦的發展與網路的普及’數位影像的儲存及傳 輸曰益重要，而且各種多媒體資訊愈來愈深入生活，不但 工作會議、通訊產品上可附加視訊影像功能，甚至汽車衛 星定位、股票交易也需要即時影像傳遞服務’爲了提高影 像傳輸的速度與品質，影像壓縮硏究對資訊產業及人類生 活之影響重大。然而，現有多數壓縮方法常無法在強調解 壓縮影像品質之餘，兼顧壓縮比以及壓縮/解壓縮時間’ 致造成應用上的瓶頸。其中又以大量網路影像資訊的傳輸 最爲常見，在網路頻寬有限的情況下’以及伴隨著上網人 數的增加，瀏覽速度因而大大的降低，資料搜尋的品質不 佳，方便性的失真。 (請先閲讀背面之注意事 本頁) .裝· 訂 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） 490973 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明（2 ) [發明欲解決之問題] 從過去電腦科學家非常重視如何節省記憶體的PC硬 體產業時代，到現在由電腦網路與多媒體所發展之新經濟 資訊時代，讓我們大家由這些新的發展，面臨了一個更重 要的問題，那就是該如何快速的傳送資料，這實在是因爲 影像壓縮之需求太過於強烈。影像壓縮之理論與技術是人 類有史人來集中心力所灌注之所在，尤其，現今的Web當 中有著越來越多與越來越複雜的各種影像圖形，在講求時 間效率的今天，網路的傳輸越來越受到重視，網路傳輸的 品質更是大家所關心的焦點。然，傳統的壓縮方式，如 Huffman編碼、Shannon-Fano編碼、算術編碼、LempelZiv 壓縮法、分頻編碼法及轉換編碼等，往往不是壓縮效率不 佳，不然就是壓縮/解壓縮程序太過煩雜，導致花很多時 間在壓縮/解壓縮過程。而現今利用類神經網路、模糊理 論及基因演算法，雖能達到較佳的壓縮率及影像品質，但 在處理的過程中，往往需要更冗長的時間才能完成。另外， 上述演算法由於數學運算複雜，大部份很難利用硬體加以 實施，不能供產業上實際的應用，使得科技資訊傳送延緩， 難以被產業所利用。 因此，本發明主要目的爲提供一以灰色關聯理論方法 爲基礎的壓縮系統與裝置，可在向量量化的編碼方式達到 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） (請先閲讀背面之注意事填寫本頁) 填·490973 Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Invention Description (/) [Technical Field of the Invention] The present invention relates to an image processing system. Specifically, the present invention is an image related to vector quantization processing technology. The compression system further explains that the present invention is a kind of using a simple gray correlation calculation method, which is particularly suitable for real-time image compression applications. Within the acceptable distortion range of the human eye, it can greatly save hundreds of millions of dollars in image storage devices. Space and can be applied to various multimedia devices. [Background of the Invention] With the development of computers and the popularization of the Internet, the storage and transmission of digital images are becoming more and more important, and various multimedia information is getting more and more deeply into life. Not only can work conferences, communication products be added with video image functions, or even cars Satellite positioning and stock trading also need real-time image delivery services. 'In order to improve the speed and quality of image transmission, image compression research has a significant impact on the information industry and human life. However, most of the existing compression methods often fail to emphasize the quality of the decompressed image while taking into account the compression ratio and compression / decompression time, which causes a bottleneck in the application. Among them, the transmission of a large amount of network image information is the most common. When the network bandwidth is limited, and as the number of Internet users increases, the browsing speed is greatly reduced, the quality of data search is poor, and the convenience is distorted. . (Please read the caution page on the back first). The size of the binding and binding paper is applicable to the Chinese National Standard (CNS) A4 (210X297 mm) 490973 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the invention ( 2) [Problems to be Solved by the Invention] From the era of the PC hardware industry where computer scientists attached great importance to how to save memory, to the new era of economic information developed by computer networks and multimedia, let us all develop from these new developments. , Facing a more important problem, that is how to quickly transfer data, this is really because the demand for image compression is too strong. The theory and technology of image compression is where human history has come to gather the central force. In particular, there are more and more various image graphics in the Web today. In terms of time efficiency today, network transmission More and more attention is being paid to the quality of network transmission. Of course, traditional compression methods, such as Huffman encoding, Shannon-Fano encoding, arithmetic encoding, LempelZiv compression, frequency division encoding, and conversion encoding, etc., are often either inefficient compression, or the compression / decompression procedures are too complicated, Lead to spending a lot of time in the compression / decompression process. Nowadays, using neural-like network, fuzzy theory and genetic algorithms can achieve better compression rate and image quality, but it often takes more time to complete in the process of processing. In addition, due to the complexity of mathematical operations, most of the above algorithms are difficult to implement with hardware and cannot be used for practical applications in the industry, which delays the transfer of scientific and technological information and is difficult to be used by the industry. Therefore, the main purpose of the present invention is to provide a compression system and device based on the gray correlation theory method, which can achieve the paper size in the vector quantization encoding method applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) (please first (Read the notes on the back and fill out this page) Fill in ·
490973 A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明（) 快速壓縮與解壓縮的效果，將是具有產業上的利用性與進 步性。 本發明的另一目的爲在網路影像傳輸的快速性下，提 供人類視覺所能接受的影像重建品質之影像壓縮系統。 本發明的再一目的係提供一種易實現的硬體架構及簡 單的軟體演算法計算，尤其是在解壓縮的處理之影像解壓 縮系統。 本發明的又一目的爲根據不同的灰色影像關聯分析， 適用於不同影像複雜度的壓縮之影像處理系統，而具新穎 性及產業上的利用性。 [發明解決問題之手段] 灰色關聯分析理論係一新興之人工智慧學門，由於灰 色關聯分析模型在大量複雜的資料分析上，具有相當精準 的關聯程度解析能力，特別適合對大量的影像資訊進行具 整體性的聚類。因此，基於此一理念及實際之需求下，本 發明乃精心硏製此一基於灰色關聯分析之影像壓縮系統， 其中本系統能在一次的影像掃描回合內，即時地將相似的 影像區塊聚爲同一類，並將各類別的影像區塊資訊加以儲 (請先閲讀背面之注意事\^填寫本頁) 填寫太 、τ490973 A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 5. Description of the invention () The effect of rapid compression and decompression will be industrially usable and progressive. Another object of the present invention is to provide an image compression system of image reconstruction quality acceptable to human vision under the fast speed of network image transmission. Still another object of the present invention is to provide an easy-to-implement hardware architecture and simple software algorithm calculation, especially an image decompression system for decompression processing. Another object of the present invention is to provide an image processing system suitable for compression of different image complexity based on different gray image correlation analysis, and has novelty and industrial applicability. [Inventing the means to solve problems] Grey relational analysis theory is an emerging artificial intelligence discipline. Because the grey relational analysis model has a fairly accurate correlation degree analysis ability on a large number of complex data analysis, it is particularly suitable for a large amount of image information. Holistic clustering. Therefore, based on this idea and practical needs, the present invention carefully crafts this image compression system based on gray correlation analysis, in which the system can instantly aggregate similar image blocks in one image scan round. For the same category, and store the image block information of each category (please read the notes on the back first \ ^ Fill this page) Fill in too, τ
本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） 經濟部智慧財產局員工消費合作社印製 490973 A7 B7 五、發明說明（屮） 存’透過聚類過程，能大幅降低儲存影像資訊的記憶體空 間及耗用的網路頻寬，並維持十分良好的還原影像品質。 更重要的是，本系統之壓縮/解壓縮速度極快，適用於即 時影像傳輸之應用，此對於資訊相關產業而言當是一大福 音。此外，本項應用灰色關聯分析配合影像壓縮系統使用 之技術，在此之前並無任何人提出過，可說是一項創舉， 且更增添了此項發明的新穎性。 [發明槪述] 本發明係爲一種利用灰色方法之影像壓縮系統，係包 括有一影像擷取裝置、一灰色關聯向量壓縮單元、一影像 儲存裝置及一灰色關聯向量解壓縮單元。其主要利用影像 擷取裝置，如數位相機、掃描器或攝影機等，自外界擷取 目的影像資料，再經由已存在於微電腦處理機中的灰色關 聯向量壓縮單元，即時地對影像資訊進行灰色關聯分析與 聚類壓縮的處理，進而產生壓縮後的資訊，並且儲存於一 影像儲存裝置，同時亦可經由Internet將影像壓縮資料傳 輸至使用者端，最後再利用微電腦處理機之灰色關聯向量 解壓縮單元進行影像還原程序’以達成快速影像壓縮的功 能。 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公爱） ------- ------- 丨訂--------- (請先閱讀背面之注意事項再填寫本頁) ^U973 A7 ---- —____B7___ 五、發明說明（/ ) 以下將對本發明之動作原理及其效果做一詳盡說明，且參 閱附呈之圖式將更爲明確。 [圖式之簡單說明] 圖一係本發明之灰色方法影像壓縮系統架構圖。 圖二係本發明之主程式流程圖。 圖三係本發明之影像區塊切割及影像序列轉換實施例。 圖四係本發明之灰色關聯分析聚類流程圖。 圖五係本發明之向量編碼簿及向量指標簿表實施例。 [圖式標號之簡單說明] (請先閱讀背面之注意事項再填寫本頁) ▼裝 ----訂ί 經 濟 部 智 慧 財 產 局 消 費 合 作 社 印 製 原始影像 1 影像擷取裝置 2 灰色關聯向量壓縮單元 3 Internet 單元 4 影像儲存裝置 5 灰色關聯向量解壓縮單元 6 還原影像 7 ^紙張尺度適用巾國國家標準（CNS)A4規格（210 x 297公釐） 490973 A7 B7 五、發明説明（厶） [發明之詳細說明] 雖然本發明將參閱含有本發明較佳實施例之所附圖式 予以充份描述，但在此描述之前應暸解熟悉本行之人士可 修改在本文中所描述之發明，同時獲致本發明之功效。因 此，須暸解以下之描述對熟悉本行技藝之人士而言爲一廣 泛之揭示，且其內容不在於限制本發明 首先請參閱圖一所示，其顯示本發明主要設施架構， 包含有一影像擷取裝置2、一灰色關聯向量壓縮單元3、一 網際網路（Internet)單兀4、一影像儲存裝置5、一灰色關 聯向量解壓縮單元6。本發明透過影像擷取裝置2，例如攝 像裝置或掃描裝置等，會把外在的原始影像1擷取進來， 經由一灰色關聯向量壓縮單元3的聚類主程式運算及處理 分析後，將產生一影像壓縮檔案，並儲存於影像儲存裝置 5或藉由網際網路單元4爲傳遞影像壓縮檔案之媒介傳送至 使用者端’最後由使用者贿灰色關聯向量解壓縮單兀6的 處理，取得還原的影像。 經濟部中央標準局員工消費合作社印製 (請先閲讀背面之注意事$填寫本頁) 圖二顯示本發明中主程式流程圖，其在開始30 1後， 灰色關聯向量壓縮單元3將自影像擷取裝置2接收到目的影 像之數位資料302，該灰色關聯向量壓縮單元3將立即對影 像資料進行區塊切割3 03，其詳細切割方式請配合參考圖 本紙張尺度適用中國國家標準（CNS ) A4規格（210X：297公釐） 490973 A7 B7 五、發明説明（7) 三，切割後的影像區塊可視爲序列，是本影像壓縮系統的 基本處理單位，接著該灰色關聯向量壓縮單元3採用灰色 關聯分析，對所有影像區塊序列進行聚類3 04程序’將會 產生向量編碼簿及向量指標簿，即爲壓縮後的資訊’其詳 細聚類演算法如圖四所示；最後，將影像壓縮資訊記錄於 影像儲存媒體5或經由網際網路單元4傳送至使用者端，當 要重建該影像，即可由該灰色關聯向量解壓縮單元6進行 查表（Look Up Table)的技術，依序將各個影像區塊成員以 所屬的影像類別取代，即可快速的取得還原影像。 其中，前述圖三影像切割的範例，在影像左上角的黑 塊即一長寬均爲4個像素點的區塊（即# = 4)，將整張影像 切割爲影像區塊後，再將各個區塊轉換爲序列，其中P 1， P2，...，P16均爲灰階値。上述亦可爲其它的像素點 區塊或lx#2，只要能適當的將整張影像完全切割。 經濟部中央標準局員工消費合作社印製 (請先閲讀背面之注意事項再填寫本頁) 接著本壓縮系統將對整張影像的所有區塊而轉換的序 列作聚類運算304做進一步說明，其中圖四詳細說明本系 統灰色關聯向量壓縮單元3採用之灰色關聯分析聚類演算 法。首先將所有影像區塊均視爲子序列，並一列列地掃描 影像，將掃描過程中第一個未分類的影像區塊視爲母序列 3 04b，使用灰色關聯分析模型求出母序列與所有子序列間 的灰色關聯係數304c，並且具體求出將母序列與所有子序 本紙張尺度適用中國國家標準（CNS ) A4規格（210x297公羡） 490973 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明（爹） 列間的灰色關聯度304d，進一步將所有求出的灰色關聯度 與一預設之門檻値作大小比較3 0 4 e ’若關聯度超過門檻値 則表示該子序列與母序列具有十分相似的灰色關聯特徵’ 可歸於同一類，同理，接著從尙未被分類（或剩餘）的所 有子序列中選取第一個影像區塊，且視爲另一母序列 3 04b，進行前述的灰色關聯度304d分析，求出另一群相似 類別的影像區塊成員，直到所有的影像區塊完全被分類。 經由上述的灰色關聯分析聚類演算法處理後，所有聚類的 母序列將組成一向量編碼簿，而該所有屬各母序列類別的 影像區塊成員，則根據所對應的母序列位於向量編碼簿的 位置，依影像區塊序列紀錄其指標號碼於一向量指標簿’ 也就是說，該向量編碼簿及向量指標簿爲影像壓縮後的資 訊。其中，該預設之門檻値可根據使用者所要求的影像還 原品質、壓縮率時壓縮速度而給定。 同時，利用向量編碼簿及向量指標簿的資訊，在灰色 關聯向量解壓縮單元6的處理中，使用查表（Look Up Table) 的技術，依序將各個影像區塊成員以所屬的影像類別取 代，即可快速的取得還原影像。 在本發明的較佳實施例中，灰色關聯度的計算依關聯 特徵如下所描述，計有一般灰色關聯分析、趨勢灰色關聯 分析及面積灰色關聯分析： 本紙張又度適用中國國家標準（CNS ) A4規格（210Χ297公釐） _ 嬅 ϋϋ· —.1— ^ϋ— fi— If· I n· ml —>_i (請先閲讀背面之注意事項再填寫本頁) 訂 490973 經濟部中央標準局員工消費合作社印製 A7 B7___ 五、發明説明（f ) 一般灰色關聯分析： 設％爲參考數列（又稱爲母數列），X,·爲比較數列（又稱爲 子數列），其中 % = k)⑴，尤〇 (2)，尤〇⑶，…〜⑷} k·⑴，A (2)，1(3),...6.0)} i = 1,2,...m 該乂與％的一般灰色關聯係數爲 t：D /1 \ Δ min + CA. max ς 〇,i(k) = —---— Α0〇,ι(^ + ζΑ0^ 前述 △〜，/⑷=卜0(众)一'·⑷I A^max = max max Δ15 〇,/(Λ) i k A^min = min min 〇,/(A:) i k 其中 △ 0，/(々)：弟A個時刻（或指標、空間）•與；c。的絕對差。 △^max: 稱爲兩級最大差，該HI严△%·(々)爲第一級最大差， 表不在 '曲線上，各相鄰點與％中各相鄰點的距 離的最大差値；rnpr^axA%#)表示在各曲線找出的 最大差m严△%·⑻的基準上，再按/ = 1，2，』找出的所 有曲線中最大差的最大差値。 本紙張尺度適用中國國家標準（CNS ) A4規格（210 X 297公釐） k = \,2,···η , ζ· = 1，2,…m k = …η , ζ· = 1，2,…m (請先閱讀背面之注意事填寫本頁)This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) Printed by the Employees ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 490973 A7 B7 V. Description of the invention (屮) The storage can greatly reduce the storage of image information through the clustering process Memory space and consumed network bandwidth, and maintain very good restored image quality. More importantly, the compression / decompression speed of this system is extremely fast, which is suitable for the application of instant image transmission. This is a great blessing for the information-related industry. In addition, this application of gray correlation analysis combined with the use of image compression systems has not been proposed by anyone before. It can be said to be an initiative and adds to the novelty of this invention. [Invention description] The present invention is an image compression system using a gray method, which includes an image capture device, a gray correlation vector compression unit, an image storage device, and a gray correlation vector decompression unit. It mainly uses image capture devices, such as digital cameras, scanners, or video cameras, to capture the target image data from the outside, and then gray correlations the image information in real time through the gray correlation vector compression unit that already exists in the microcomputer processor. Analysis and cluster compression processing, and then generate compressed information and store it in an image storage device. At the same time, the image compression data can be transmitted to the user end via the Internet. Finally, the gray correlation vector is used to decompress the computer. The unit performs the image restoration process to achieve the function of fast image compression. This paper size applies to China National Standard (CNS) A4 specification (210 X 297 public love) ------- ------- 丨 order --------- (Please read the Please fill in this page again for the matters needing attention) ^ U973 A7 ---- —____ B7___ V. Description of the Invention (/) The operation principle and effect of the present invention will be described in detail below, and it will be more clear with reference to the attached drawings. [Brief description of the drawings] FIG. 1 is a structural diagram of a gray method image compression system of the present invention. Figure 2 is a flowchart of the main program of the present invention. FIG. 3 is an embodiment of image block cutting and image sequence conversion according to the present invention. FIG. 4 is a grey relational analysis clustering flowchart of the present invention. FIG. 5 is an embodiment of a vector coding book and a vector index book table of the present invention. [Simplified description of figure numbers] (Please read the precautions on the back before filling out this page) ▼ Installation ---- Order the original image printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 1 Image capture device 2 Gray correlation vector compression Unit 3 Internet Unit 4 Image Storage Device 5 Gray Correlation Vector Decompression Unit 6 Reconstructed Image 7 ^ Paper Size Applicable National Standard (CNS) A4 Specification (210 x 297 mm) 490973 A7 B7 V. Description of Invention (厶) [ Detailed description of the invention] Although the present invention will be fully described with reference to the accompanying drawings containing preferred embodiments of the present invention, it should be understood that those skilled in the art can modify the invention described herein before describing this, and The effect of the present invention is obtained. Therefore, it must be understood that the following description is a broad disclosure for those familiar with the skills of the bank, and its content is not intended to limit the present invention. First, please refer to FIG. 1, which shows the main facility architecture of the present invention, including an image capture. A fetching device 2, a gray correlation vector compression unit 3, an Internet unit 4, an image storage device 5, and a gray correlation vector decompression unit 6. According to the present invention, an external original image 1 is captured through an image capturing device 2 such as a camera device or a scanning device. After being calculated and processed by a clustering main program of a gray correlation vector compression unit 3, An image compression file is stored in the image storage device 5 or transmitted to the user terminal through the Internet unit 4 as a medium for transmitting the image compression file. Finally, the user deducts the unit 5 from the gray correlation vector to obtain the processing. The restored image. Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs (please read the note on the back page to fill out this page). Figure 2 shows the main program flow chart of the present invention. After starting 30 1, the gray correlation vector compression unit 3 will self-image The capture device 2 receives the digital data 302 of the target image. The gray correlation vector compression unit 3 will immediately cut the image data into blocks 03 03. For detailed cutting methods, please refer to the drawing. The paper dimensions are applicable to the Chinese National Standard (CNS) A4 specifications (210X: 297 mm) 490973 A7 B7 V. Description of the invention (7) Third, the cut image block can be regarded as a sequence, which is the basic processing unit of this image compression system. Then the gray correlation vector compression unit 3 uses Grey relational analysis, clustering all image block sequences. The 3 04 program 'will generate a vector codebook and a vector index book, which is the compressed information'. The detailed clustering algorithm is shown in Figure 4. Finally, the The image compression information is recorded in the image storage medium 5 or transmitted to the user end via the Internet unit 4. When the image is to be reconstructed, the gray scale 6 decompression unit vector technology lookup table (Look Up Table), the respective image block sequentially member belongs to the image category substituted, to quickly restore the image acquisition. Among them, in the example of image cutting shown in FIG. 3, the black block in the upper left corner of the image is a block with a length and width of 4 pixels (that is, # = 4). After cutting the entire image into image blocks, Each block is converted into a sequence, where P 1, P2, ..., P16 are all gray-scale 値. The above may also be other pixel blocks or lx # 2, as long as the entire image can be completely cut properly. Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs (please read the notes on the back before filling this page). Then the compression system will further perform the clustering operation 304 on the sequence converted from all blocks of the entire image. Figure 4 illustrates the grey correlation analysis clustering algorithm used in the grey correlation vector compression unit 3 of the system. First, all image blocks are considered as sub-sequences, and the images are scanned line by line. The first unclassified image block in the scanning process is regarded as the mother sequence 3 04b. The gray sequence analysis model is used to find the mother sequence and all The gray correlation coefficient 304c between the subsequences, and the paper size of the parent sequence and all the subsequences are specifically calculated to apply the Chinese National Standard (CNS) A4 specification (210x297 public envy) 490973 Printed by the Consumers Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the invention (Daddy) The gray correlation between the columns is 304d, and further compare all the gray correlations obtained with a preset threshold 値 3 0 4 e 'If the correlation exceeds the threshold 表示, it indicates the subsequence The gray correlation features that are very similar to the parent sequence can be classified into the same class, and the same reason, then the first image block is selected from all sub-sequences that are not classified (or remaining), and regarded as another parent sequence 3 04b, perform the foregoing gray correlation 304d analysis to obtain another group of image block members of a similar category until all the image blocks are completely classified. After processing by the above-mentioned gray correlation analysis clustering algorithm, the mother sequences of all clusters will form a vector codebook, and all the image block members belonging to each mother sequence category will be located in the vector code according to the corresponding mother sequences. The position of the book is recorded in a vector index book according to the image block sequence. That is, the vector encoding book and the vector index book are information after image compression. The preset threshold 値 can be set according to the image restoration quality and compression speed required by the user. At the same time, using the information of the vector encoding book and the vector index book, in the processing of the gray correlation vector decompression unit 6, a look-up table technology is used to sequentially replace each image block member with the image category to which it belongs , You can quickly get the restored image. In the preferred embodiment of the present invention, the calculation of the gray correlation degree according to the correlation characteristics is described as follows, including general gray correlation analysis, trend gray correlation analysis, and area gray correlation analysis: This paper is also applicable to the Chinese National Standard (CNS) A4 specifications (210 × 297 mm) _ 嬅 ϋϋ · —.1— ^ ϋ— fi— If · I n · ml — > _i (Please read the notes on the back before filling this page) Order 490973 Central Standards Bureau of the Ministry of Economic Affairs Printed by employee consumer cooperative A7 B7___ V. Description of the invention (f) General gray correlation analysis: Let% be the reference sequence (also known as the parent sequence), and X, · be the comparison sequence (also called the child sequence), where% = k ) ⑴, especially 〇 (2), especially 〇⑶, ... ~ ⑷} k · ⑴, A (2), 1 (3), ... 6.0)} i = 1,2, ... m The general gray correlation coefficient of% is t: D / 1 \ Δ min + CA. max ς 〇, i (k) = —---— Α0〇, ι (^ + ζΑ0 ^ The aforementioned △ ~, / ⑷ = 卜 0 (Many) One '· ⑷I A ^ max = max max Δ15 〇, / (Λ) ik A ^ min = min min 〇, / (A :) ik where △ 0, / (々): brother A moment (or Indicators, space) • and; c. Absolute difference. △ ^ max: It is called the two-level maximum difference, and the HI is strictly △% · (々) is the first-level maximum difference. The table is not on the 'curve. ; Rnpr ^ axA% #) represents the maximum difference 严 of the maximum difference among all the curves found on the basis of the maximum difference m strict △% · ⑻ found on each curve, and then press / = 1,2. This paper size applies Chinese National Standard (CNS) A4 specification (210 X 297 mm) k = \, 2, ·· η, ζ · = 1, 2, ... mk =… η, ζ · = 1, 2, … M (Please read the notes on the back and fill out this page)
、1T 4 490973 A7 B7 五、發明説明（丨汐） △ min: 爲兩級最小差，意義與兩級最大差相似。 《·· 稱其爲分辨係數，一般（是分佈於〇〜1之實數， 其具體取値視實際情況而定。 因此，可求出；^•與對應X。的一般灰色關聯度 / W “。，xz·)=| Ζ:=ι 々 Ά) 接著，在趨勢灰色關聯分析與面積灰色關聯分析中，因與 一般灰色關聯分析相似，故僅將其不同測度公式說明如 下。 趨勢灰色關聯分析： 該七與々的趨勢灰色關聯係數爲 (請先閱讀背面之注意事填寫本頁)1T 4 490973 A7 B7 V. Description of the invention (丨 xi) △ min: It is the minimum difference between two levels, and its meaning is similar to the maximum difference between two levels. "·· It is called the resolution coefficient, generally (is a real number distributed from 0 to 1, which depends on the actual situation. Therefore, it can be obtained; ^ • the general gray correlation with the corresponding X / W" ，, xz ·) = | ZZ: = ι 々Ά) Then, in the trend gray correlation analysis and area gray correlation analysis, because it is similar to the general gray correlation analysis, only the different measurement formulas are explained below. Trend gray correlation analysis : The grey correlation coefficient of the trend of the seven and 々 is (Please read the note on the back first and fill in this page)
、τ ξ8^), Τ ξ8 ^)
Asoj(k) + CA\ k = \，2，···η， ί = 1,2,. 經濟部中央標準局員工消費合作社印製 \與對應χ。的趨勢灰色關聯度爲 rW(w)=忐 〇5。，办) 1，2” 其中 Δ5 〇,/ (k) = |χ〇 (A: +1) - ^c〇 (k) - (xt (k+ 1)- xi (Λ:))| 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） 490973 A7 B7 五、發明説明（"） A5max = max max Δ5〇,/(A:) i k A5min =minminA5〇,/(A:) i k 面積灰色關聯分析： 該&與％的面積灰色關聯係數爲 众=l，2”“w -1， / = 1，2，· ζΑ\Asoj (k) + CA \ k = \, 2, ··· η, ί = 1,2 ,. Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs \ and corresponding χ. The trend gray correlation is rW (w) = 忐 〇5. , Office) 1, 2 ”where Δ5 〇, / (k) = | χ〇 (A: +1)-^ c〇 (k)-(xt (k + 1)-xi (Λ :)) | Applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) 490973 A7 B7 V. Description of the invention A5max = max max Δ5〇, / (A :) ik A5min = minminA5〇, / (A :) ik area Gray correlation analysis: The gray correlation coefficient of the area of the & and% is mode = 1, 2 "" w -1, / = 1, 2, · ζΑ \
A\j(k)-¥CAA A: = 1，2”··《 — 1， / = 1，2，· 七與對應；c。的趨勢灰色關聯度爲 〆〇，，= 〆（々，'·)= n-l I (卜0 (灸+1) -、·(仓+1)1+卜0⑷-七⑷|)， if (^0 + (k + \))(x0 (k) - X, (k)) > 0 (請先閲讀背面之注意事填寫本頁) β .裝· 訂 △\办) (^ + 1) - ^ + 1))2+(χ0(Λ)4(Α〇)2 2(卜。(A：+1) 一 a (灸+1)|+卜。⑷一〜⑷I) ’ if (^0 (^ + 1)- (k + l))(x0 (k) - x. (k)) > 0 m 經濟部中央標準局員工消費合作社印製 A^max =maxmaxA>i〇,i(A：) i k △^min = min min zVV, (A：) 經過詳細說明本發明的較佳實施例之後，熟悉該項技 術人士可淸楚的瞭解，並在不脫離下述申請專利範圍與精 神下可進行各種變化與改變，而且本發明亦不受限於說明 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） 竹 υ973 A7 B7 五、發明說明（丨>) 書之實施例的實施方式。 [發明之功效] (請先閱讀背面之注意事項再填寫本頁) 根據本發明所實施的灰色方法之影像處理系統與裝 置’將可有效地改善習知技術在影像壓縮缺失，並可獲致 以下的功效： 1. 由於只需一次（One Pass)的影像聚類過程，故可加速壓 縮程序。 2. 本發明之影像壓縮系統，可對不同複雜的影像進行最 佳的壓縮品質，其中一般灰色關聯量測適用於變化和 緩的影像，例如人像、風景等；趨勢灰色關聯量測適 用於變化劇烈的影像，如電路面板影像；面積灰色關 聯量測適用於有雜訊的影像處理，可以減少影像壓縮 對雜訊的敏感度。 3. 軟體演算法計算簡單，在解壓縮的過程，硬體架構容 易實現。 經濟部智慧財產局員工消費合作社印製 4. 可以針對多張相似的影像進行壓縮，產生一共同的編 碼簿，即可大大減低壓縮後的資料量及傳送時間。 5 . —般的影像，在8x8的影像區塊分割壓縮後，普遍能 獲得人眼可接受的視覺品質，約可達到60倍的壓縮效 果。 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公爱） 490973 A7 B7 五、發明說明（丨> ) 綜合言之，本發明計算之方法簡單，壓縮之時間相當 快速，已符合於新穎性及實用性之要件，在專利申請前從 未有相同或類似之物品被核准專利或公開使用過，且其可 有效地應用在各種多媒體的影像處理場合，大大節省了影 像傳輸的時間，並且提升了影像重建的品質，極具實用價 値’當符合專利法之規定，茲依法提出專利申請，敬請貴 審查委員明鑑，並賜准專利，實感德便。 emt 1 i ·ϋ ϋ 1 ϋ ϋ ϋ ^1 ^1 ϋ _1 ϋ 一-0, I ·ϋ ϋ -ϋ ϋ ϋ I ^1 I (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐）A \ j (k)-¥ CAA A: = 1, 2 "··" — 1, / = 1, 2, 7 · Correspondence; c. The trend gray correlation degree is 〆〇 ,, = 〆 (々, '·) = Nl I (Bu0 (Moxibustion + 1)-, · (Cang + 1) 1 + Bu0⑷- 七 ⑷ |), if (^ 0 + (k + \)) (x0 (k)-X , (k)) > 0 (Please read the notes on the back and fill out this page first) β. Binding and ordering △ \ Office) (^ + 1)-^ + 1)) 2+ (χ0 (Λ) 4 (Α 〇) 2 2 (Bu. (A: +1) a (moxibustion +1) | + Bu. ⑷ 一 ~ ⑷I) 'if (^ 0 (^ + 1)-(k + l)) (x0 (k )-x. (k)) > 0 m Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs A ^ max = maxmaxA > i〇, i (A :) ik △ ^ min = min min zVV, (A :) After After explaining the preferred embodiment of the present invention in detail, those skilled in the art can understand it well, and can make various changes and modifications without departing from the scope and spirit of the patent application described below, and the present invention is not limited to the description This paper size applies to Chinese National Standard (CNS) A4 (210X297 mm) Bamboo 973 A7 B7 V. Description of the invention (丨 >) The embodiment of the book. [Effect of the invention] (Please read the note on the back first Refill (Write this page) The image processing system and device of the gray method implemented according to the present invention will effectively improve the lack of conventional techniques in image compression, and can achieve the following effects: 1. Because only one (One Pass) The image clustering process can speed up the compression process. 2. The image compression system of the present invention can perform the best compression quality on different complex images, of which the general gray correlation measurement is suitable for changing and gentle images, such as portraits, landscapes, etc. Etc .; trend gray correlation measurement is suitable for images that change rapidly, such as circuit panel images; area gray correlation measurement is suitable for image processing with noise, which can reduce the sensitivity of image compression to noise. 3. Software algorithm calculation It is simple, and the hardware architecture is easy to implement in the process of decompression. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 4. It can compress multiple similar images to generate a common codebook, which can greatly reduce the compression. The amount of data and the transmission time. 5 --- ordinary image, after 8x8 image block segmentation and compression, can generally obtain human eyes Accepted visual quality can reach about 60 times the compression effect. This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 public love) 490973 A7 B7 V. Description of the invention (丨 >) In summary, this The calculation method of the invention is simple, and the compression time is quite fast. It has met the requirements of novelty and practicability. Before the patent application, no identical or similar article has been approved for patent or public use, and it can be effectively used in Various multimedia image processing occasions greatly save the time of image transmission and improve the quality of image reconstruction. It is of great practical value. When the patent law is complied with, patent applications are submitted in accordance with the law. Your reviewers will be informed and will give you Quasi-patent, real sense. emt 1 i · ϋ ϋ 1 ϋ ϋ ϋ ^ 1 ^ 1 ϋ _1 ϋ 一 -0, I The paper size printed by the Bureau ’s Consumer Cooperatives applies the Chinese National Standard (CNS) A4 (210 X 297 mm)