基于GEP多标记学习的图像超分辨率复原算法

基于图像单一特征进行支持向量机预分类的超分辨率复原算法通过离线建立分类模型,减少了传统基于范例学习算法的样本块误匹配现象,提高了图像质量和计算速度。但由于图像特征的多样性,此类算法易造成复原结果的不稳定。本文提出一种基于基因表达式编程多标记学习的超分辨率复原算法,筛选出与目标图像相关性高的样本子库,在多标记框架下进行样本预分类。实验结果表明,本文算法稳定性强、鲁棒性好,进一步缩小了低分辨率图像块的匹配范围,更好的提高了超分辨率复原的效果和效率。

Image Super-resolution algorithm based on GEP multi-label learning

The SVM pre-classified super-resolution algorithm is based on single image feature and builds off-line disaggregated models. It reduces the mis-matching of tranditional example-based restoration algorithms, improves the image quality and running speed. However, the SVM-based algorithm easily leads to unstable recovered results because of the diversity of image features. For such problems, we propose a super-resolution restoration algorithm based on GEP multi-label learning. The algorithm selects the subset of sample library which is highly related to the object image and pre-classifies the image samples under the multi-label framework. The experimental results show that the proposed method is robust and stable. Specifically, the algorithm further reduces the matching range of low resolution image blocks and promotes the restoration effectiveness and efficiency.