Image Reconstruction Based on Compressive Sensing With Equalization Quantization Noise Model and Multi-direction Model

An error estimate method based on equalization quantization noise model was proposed for nonuniform quantization noise in image codec. With designed equalization matrix,a norm constraint which could enhance the quality of CS recovery significantly was shown. Furthermore,for directional textures of images,a multi-direction operator which could enhance the quality of CS optimize recovery was shown. A CS-based JPEG decoding scheme based on quantization error estimate and multi-direction operator was also presented,and experimental evidence exhibits more gains over CS reconstruction without error estimation and multi-direction operator.