New SVD image compression algorithm based on salient object detection
一种基于感兴趣区域提取的新型SVD图像压缩算法

A new image-coding algorithm (SODSVD) that combined the standard Singular Value Decomposition (SVD) algorithm and the salient object detection technique was proposed. During encoding the images, the salient object detection was used to split the source image into adaptive smaller blocks based on the complexity and visual attention, while smaller blocks were transformed by use of SVD. A simple strategy was proposed to split the source images. Simulation results show that the new algorithm provides good distortion, bit rate and image quality.