| 基于视觉注意计算模型的图像压缩新方法 |
| |
| 引用本文： | 刘伟,汪云九,童勤业.基于视觉注意计算模型的图像压缩新方法[J].科技通报,2006,22(6):775-780. |
| |
| 作者姓名： | 刘伟 汪云九 童勤业 |
| |
| 作者单位： | 1. 浙江大学,生物医学工程与仪器科学学院,杭州,310027
2. 中国科学院,生物物理研究所,北京,100008 |
| |
| 摘 要： | 图像可以被分为重要的感兴趣区域和不重要的背景区域。提出了一种新的基于感兴趣区域的图像压缩方法。该方法首先使用一个基于视觉生理和心理物理实验基础的视觉注意计算模型计算图像中的感兴趣区域,然后用JPEG算法对感兴趣区域和背景区域采用不同的压缩比进行压缩。实验结果表明,该算法具有比JPEG算法更高的压缩效率,同时保证了图像良好的视觉效果。
|
| 关 键 词： | 视觉注意 感兴趣区域 图像压缩 特征抽取 |
| 文章编号： | 1001-7119（2006）06-0775-06 |
| 收稿时间： | 2005-03-06 |
| 修稿时间： | 2005年3月6日 |
A New Image Compression Approach Based on Visual Attention Computational Model |
| |
| LIU Wei,WANG Yun-Jiu,TONG Qin-Ye.A New Image Compression Approach Based on Visual Attention Computational Model[J].Bulletin of Science and Technology,2006,22(6):775-780. |
| |
| Authors: | LIU Wei WANG Yun-Jiu TONG Qin-Ye |
| |
| Institution: | 1. College of Biomedical Engineering and Instrument Science ,Zhejiang University, HangZhou 310027, China; 2. Institute of Biophysics, Chinese Academy of Sciences, Beijing 100008, China |
| |
| Abstract: | An image can be segmented into the region of interesting(ROI),which is considered important,and the back-ground,which is less important.In this paper,a new ROI based image compression algorithm is proposed.The first step of the algorithm is to find out the ROI in the image.A saliency-based bottom-up visual attention computational model which is motivated by visual physiological and psychophysical experimental results is used.The second step is encoding and which is based on the JPEG algorithm.The ROI of the image is compressed with a low compression ratio and the back ground with a high one.The reconstruction algorithm of the compressed image is like that of the JPEG algorithm.The al-gorithm proposed has higher compression ratio than JPEG algorithm and the image compressed has a perceptually high quality.We test our algorithm with lots of natural images and satisfy with the result. |
| |
| Keywords: | visual attention region of interesting(ROI) image compression feature extraction |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
