摘要
图像质量评价是图像处理领域的一个研究热点。文中将视觉注意模型结合到传统结构相似度(SSIM)方法中,尝试将视觉注意机制的研究成果引入到图像质量评价领域,以获取更好的评价效果。新方法采用Itti视觉注意计算模型提取图像的全局显著图,然后用中介真值程度(MMTD)理论确定各局部窗口的权值,在评价过程中对显著程度高的区域给予更高的权重。实验结果表明改进后的MMTD-SSIM算法在感兴趣区域突出的图像质量评价中较传统SSIM算法更加准确有效,更加接近人类视觉的主观评价。
Image quality assessment is a hot research topic in the field of image processing. In this paper,the study of the visual attention mechanism is introduced into the image quality assessment through combination of visual attention model into SSIM,trying to obtain bet-ter evaluation result. The new method uses Itti visual attention model to extract saliency map,then determines local weights with the theo-ry of MMTD,in the process of evaluation of salient area for a higher weight. The experiment results show that MMTD-SSIM algorithm is more accurate and effective than SSIM in the image quality assessment of image with outstanding region of interest,and more close to the human visual subjective assessment.
出处
《计算机技术与发展》
2015年第1期226-229,共4页
Computer Technology and Development
基金
国家自然科学基金资助项目(61170322)
关键词
图像质量评价
视觉注意模型
显著图
结构相似度
中介真值程度
Image Quality Assessment（IQA） visual attention model saliency map Structural Similarity Image Measurement（SSIM） Measure of Medium Truth Degree（MMTD）
