一种基于视觉熵的图像分割压缩方法

基金项目:

国家自然科学基金资助课题(No.69873012)

详细信息

Image Compression with Visual Entropy-Based Segmentation

摘要: 基于视觉熵概念提出了一种静止图像分割压缩方法.通过对人类视觉系统特性的归纳,总结了基于视觉熵的图像分割原理,提出了用于量化图像特征的数学定义和基于视觉嫡的分割算法.实验结果表明,这种基于视觉熵的图像分割压缩算法既提高了压缩比,又能保证压缩后的重建图像整体上具有高的主观视觉感知质量.
- 视觉熵 /
- 图像分割 /
- 图像编码
Abstract: An image compression method with Visual Entropy-Based segmentation is presented. Firstly the induction of the characteristics of Human Vision System(HVS) and the principles for Visual Entropy-based segmentation is made. Then the mathematic definition for quantification of image character and the algorithm for Visual Entropy-Based segmentation are fully described. The experimental results have shown that image compression with Visual Entropy-Based segmentation can not only gets a rather low bit rate but also gives satisfactory subjective perceptual quality. This method has well emulated the properties of HVS.
- visual entropy (VE) /
- image segmentation /
- image coding