摘要
针对当前深度超分网络参数量非常庞大,且推理速度缓慢,而轻量化网络在复杂环境条件下无法表达图像深层特征的问题,提出基于特征增强的轻量化可逆超分辨率网络.首先提出边缘特征残差,配合所提的边缘相似损失指导模型重建,增强重建图像对纹理轮廓的表达能力;然后补充新的小波特征核,使小波变换支持任意缩放因子的重建任务;最后引入全局特征提取模块,在特征图中嵌入自注意力机制,提取全局特征.在缩放因子为4时的基准测试集Set5上的实验结果表明,与SwinIR-light相比,所提网络表现更优, PSNR提升0.41 dB,参数量减少244k,推理时间缩短49.05%.
Deep super-resolution reconstruction networks have a large number of parameters and slow inference speed.Lightweight networks unable to express deep features of images under complex environmental conditions.Aiming at those issues,a lightweight reversible super-resolution network based on feature enhancement is proposed.Firstly,edge feature residuals are proposed,combined with the proposed edge similarity loss to guide model reconstruction and enhance the expression of texture contours.Then,add a new wavelet feature kernel to support the reconstruction task with any scaling factors.Finally,a global feature extraction module is introduced to embed a self attention mechanism in the feature map to extract global features.The proposed network showed better performance than SwinIR-light on the benchmark test Set5 with a scaling factor of 4.It improved the PSNR by 0.41 dB,decreased the number of parameters by 244k,and reduced the inference time by 49.05%.
作者
杨本臣
李浩然
金海波
Yang Benchen;Li Haoran;Jin Haibo(School of Software,Liaoning Technical University,Huludao 125000)
出处
《计算机辅助设计与图形学学报》
北大核心
2025年第4期668-677,共10页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(62173171)。
关键词
超分辨率
小波变换
边缘特征
可逆神经网络
super-resolution
wavelet transform
edge features
reversible neural network
