采用稀疏表示和小波变换的超分辨率重建算法

为了提高超分辨率重建图像的质量,提出一种基于稀疏表示和小波变换的超分辨率重建算法.首先,将小波变换的多尺度性、多方向性与稀疏表示的灵活性相结合,构建一种双稀疏编码(DSC)模型,提高稀疏系数的精度.然后,在双稀疏编码模型中引入局部线性嵌入正则化项(LLER),以更好地保留图像的结构;在重建过程中,对输入的低分辨率图像进行小波分解,得到3幅不同方向的高频子图,并采用提出的模型对其进行重建.最后,利用逆小波得到最终的高分辨率图像.实验结果表明:与多种主流的超分辨率算法相比,文中算法无论在主观视觉效果还是在峰值信噪比和结构相似度两个客观评价指标上,都取得了更好的效果.

Super-Resolution Reconstruction Algorithm Using Sparse Representation and Wavelet Transform

To improve the quality of super-resolution reconstruction images, a super-resolution reconstruction algorithm based on sparse representation and wavelet transform is proposed. Firstly, combining the multi-scale and multi-directionality of wavelet transform with the flexibility of sparse representation, a dual sparse coding(DSC)model is constructed to improve the accuracy of sparse coefficients. Then, a locally linear embedding regularization(LLER)term is introduced to better preserve the structure of the image. In the process of image reconstruction, three high-frequency subbands with different directional characteristics, obtained by performing wavelet decomposition on the input LR image, are reconstructed by the proposed LLER-DSC model, respectively. Finally, the final high-resolution image is obtained by inverse wavelet transform. Experiments illustrate that the proposed approach outperforms several state-of-the-art super-resolution algorithms in terms of subjective visual quality and objective evaluation indices including peak signal-to-noise ratio and structural smilarity.