摘要
针对快速相关矢量量化算法重建图像质量不高、存在明显方块效应的问题,采用后向搜索一定区域的预测与有限状态矢量量化相结合的编码方法,并用MEENS-2和PDS结合搜索算法完成码字搜索。实验结果表明,该编码方法既提供了高压缩比,又减少了额外失真的影响,保证了图像的重建质量。
Fast Correlation Vector Quantization（BFCVQ） have the shortage in degraded quality of rebuilding image and block artifact, against which, an image encoding algorithm is presented in this paper which integrates exploiting a finite region backward to predict the input block with the finite-state vector quantization（FSVQ）. The Improvement of Equal-average Equal-variance Nearest Neighbor Search algorithm （IEENNS-2） and Partial Distortion Search algorithm （PDS） are combined to search codewords. Experimental resuits show that the proposed algorithm provides high compression ratio, reduced the additional distortion and have better rebuilding quality.
出处
《微计算机应用》
2007年第3期236-240,共5页
Microcomputer Applications
关键词
矢量量化
相关预测
图像编码
Vector Quantization（ VQ）, correlation predictive, image encoding
