电子测量与仪器学报

胡雨航,赵磊,李恒,刘辉.多特征选择与双向残差融合的无监督水下图像增强[J].电子测量与仪器学报,2023,37(9):190-202

多特征选择与双向残差融合的无监督水下图像增强

Unsupervised underwater image enhancement with multi-feature selection and bidirectional residual fusion

DOI：

中文关键词: 无监督模型循环生成对抗网络多特征选择双向残差融合水下图像增强

英文关键词:unsupervised models recurrent generative adversarial networks multi-feature selection bidirectional residual fusion underwater image enhancement

基金项目:国家自然科学基金项目(62263016)、云南省科技厅面上项目(202001AT070038)资助

作者	单位
胡雨航	1.昆明理工大学信息工程与自动化学院
赵磊	1.昆明理工大学信息工程与自动化学院
李恒	1.昆明理工大学信息工程与自动化学院
刘辉	1.昆明理工大学信息工程与自动化学院

Author	Institution
Hu Yuhang	1.Faculty of Information Engineering and Automation, Kunming University of Science and Technology
Zhao Lei	1.Faculty of Information Engineering and Automation, Kunming University of Science and Technology
Li Heng	1.Faculty of Information Engineering and Automation, Kunming University of Science and Technology
Liu Hui	1.Faculty of Information Engineering and Automation, Kunming University of Science and Technology

摘要点击次数: 612

全文下载次数: 782

中文摘要:

如今,利用合成的成对数据集训练的有监督模型泛化能力弱,在多变的实际水下环境中表现不佳,而无监督模型虽摆脱了成对数据集的依赖,但生成图像可能因缺少特征信息导致图像视觉质量较差。故以循环生成对抗网络为架构,提出多特征选择与双向残差融合的水下图像增强方法。一方面,设计以混合注意力为基础的多特征选择模块对水下图像的多种特征进行选择,再由双向残差融合对传统 U 型跳跃连接进行优化,使图像特征高效表达,有效恢复水下图像的纹理与色彩。另一方面,在判别器中引入混合注意力并提出内容感知损失和风格感知损失,保证增强图像在全局内容、局部纹理、风格特征等方面和清晰图像一致。与现有的无监督和有监督模型相比较,该模型 PSNR 分别提高了 6%和 2%,SSIM 分别提高了 4%和 3%,对水下图像有着显著的增强效果,在色彩真实度和饱和度上相比其他现有方法更加优秀。

英文摘要:

Currently, the supervised models trained on synthetic paired datasets have weak generalization ability and perform poorly in diverse real underwater environments. Although unsupervised models are not dependent on paired datasets, the lack of feature information may result in the generated images with poor visual quality. Therefore, with the architecture of cyclic generation adversarial networks, the underwater image enhancement method of multi-feature selection and bidirectional residual fusion is proposed. On one hand, a multi-feature selection module based on mixed attention is designed to select multiple features of underwater images. Furthermore, the bidirectional residual fusion is used to optimize traditional U-shaped skip connection, which realizes high-efficiency expression of image features and effectively restores the texture and color of underwater images. In addition, mixed attention is introduced and content-aware loss and style-aware loss are proposed in the discriminator to ensure that the enhanced image is consistent with the clear image in terms of global content, local texture, and style features. The PSNR of the proposed model is improved by 6% and 2%, respectively, compared with the existing unsupervised and supervised models. Additionally, SSIM is improved by 4% and 3%, respectively. With a significant enhancement effect on underwater images, the proposed method demonstrates superiority over other existing methods in terms of color fidelity and saturation.

查看全文查看/发表评论下载PDF阅读器