該系統的完整架構概分為三大模組：(1)影像前處理模組，(2)特徵能量運

該系統的完整架構概分為三大模組：(1)影像前處理模組，(2)特徵能量運算模組，以及(3)多層級壓縮模組。其中影像前處理模組首先針對一張輸入的數位化影像進行背景區域分割、梯度增強及過濾背景雜訊。接著，特徵能量運算模組以我們所發展的「固定特徵能量最小化法」 (Fixed Feature Energy Minimization，FFEM)進行特徵能量強度運算，以辨識影像中重要組織存在的可能性，並據以進行後續的分級壓縮。最後的多層級壓縮模組主要利用第二個模組運算所得的特徵能量，進一步將醫學影像區塊劃分為不同等級，再以JPEG2000施予相對壓縮率的分級壓縮，並以此資料進行儲存與傳輸。

結果 (英文) 1:

Full institution of this system are summarized into three large mode group: (1) finish die-given before the image, (2) features energy chasing after a die integral, and (3) multiple coating class integral of pressure product can die. Images before I do finish die-given 針 Yes Zhang input of digital progress row background area type of image segmentation CDT, gradient and background 雜 訊 over 濾. Then, in characteristic energy chasing after a die integral to our release show "fixed features of energy minimization method" (Fixed Feature Energy Minimization,FFEM) features energy intensity also set the progress line count, with the possibility of differentiation of important team issue exists in the main image, and data to 續 after progress line bug classification pressure product can. The last coating of class integral of pressure group mainly use the second one of this product can die die chasing after features available to calculate energy progress step advises Medical imaging area 塊 drawn into different ASTM c, again with JPEG2000 giving Yes pressure bug classification of this product can rate pressure product can, does this information material progress line 儲 survival stub 傳 input.

結果 (英文) 2:

Almost complete architecture of the system is divided into three modules: (1) image pre-processing module, (2) the characteristic energy calculation module, and (3) multi-stage compression module. Wherein the image processing module before the first digitized input image for a background region segmentation, enhancement, and filtering gradient background noise. Then, with characteristic energy calculation module we developed a "fixed feature energy minimization method" (Fixed Feature Energy Minimization, FFEM) to characterize the energy intensity of operation, the possibility of an important organization to recognize the existence of the image, and data for subsequent Compression grading. Finally, multi-stage compression module mainly use the second module operation resulting characteristic energy, medical imaging blocks further divided into different grades, and then administered to grade JPEG2000 compression relative compression ratio, and thus data for storage and transmission.

結果 (英文) 3:

The whole architecture of the system is divided into three modules: (1) image pre-processing module, (2) feature energy computing module, and (3) multilayer compression module. The image preprocessing module first for a piece of input digital images of regional background segmentation, gradient enhancement and noise filtering. And feature energy computation module to our development of the "fixed features energy minimization method (fixed feature energy minimization, ffem) are characteristics of energy intensity of operation, in order to distinguish the possibility of an important organization in knowledge of image, and according to the grade of the subsequent compression.Finally, the multi layer compression module mainly uses second modules to further divide the medical image blocks into different grades, and then JPEG2000 is applied to the classification of the relative compression rate, and the data is stored and transmitted.