电子设计工程
電子設計工程
전자설계공정
ELECTRONIC DESIGN ENGINEERING
2014年
15期
78-81
,共4页
离散曲波变换算法%曲波加权算法%图像识别%识别率
離散麯波變換算法%麯波加權算法%圖像識彆%識彆率
리산곡파변환산법%곡파가권산법%도상식별%식별솔
discrete curvelet transform algorithm%curvelet weighted algorithm%image recognition%recognition rate
识别率的提升是图像处理技术的关键环节，笔者针对第二代曲波变换算法在图像识别处理过程中，所存在的图像边缘“振铃”效应和由于“楔形基”的特性所导致的图像失真问题，提出了第二代曲波加权改进算法及对图像识别的实现过程，并且分别通过ORL和Yale图像进行了对比仿真实验，证明了较传统的小波加权双向二维主成分分析算法在对识别中有明显的提高，从而验证了该算法在图像识别处理上的可行性和有效性。
識彆率的提升是圖像處理技術的關鍵環節，筆者針對第二代麯波變換算法在圖像識彆處理過程中，所存在的圖像邊緣“振鈴”效應和由于“楔形基”的特性所導緻的圖像失真問題，提齣瞭第二代麯波加權改進算法及對圖像識彆的實現過程，併且分彆通過ORL和Yale圖像進行瞭對比倣真實驗，證明瞭較傳統的小波加權雙嚮二維主成分分析算法在對識彆中有明顯的提高，從而驗證瞭該算法在圖像識彆處理上的可行性和有效性。
식별솔적제승시도상처리기술적관건배절，필자침대제이대곡파변환산법재도상식별처리과정중，소존재적도상변연“진령”효응화유우“설형기”적특성소도치적도상실진문제，제출료제이대곡파가권개진산법급대도상식별적실현과정，병차분별통과ORL화Yale도상진행료대비방진실험，증명료교전통적소파가권쌍향이유주성분분석산법재대식별중유명현적제고，종이험증료해산법재도상식별처리상적가행성화유효성。
The ascension of recognition rate is a key link in the process of image processing technology. Aimed at the second generation of curvelet transform algorithm in the process of image recognition processing, the existing image edge “ringing”effect and image distortion as a result of the nature of the “wedge base”, the second generation curvelet weighted improved algorithm is proposed and the implementation process of image recognition. And respectively compared the ORL with Yale image simulation experiment, proves that it is better in the recognition than traditional wavelet weighted bi-directional two-dimensional principal component analysis algorithm, so as to verify the feasibility and effectiveness of the algorithm on image recognition processing.
