Compressive sensing inspired self-supervised single-pixel imaging
受压缩感知启发的自监督单像素成像
机构 * Institute of Artificial Intelligence (TeleAI), China Telecom, China(中国电信人工智能研究院(TeleAI)) ; School of Integrated Circuits, Shanghai Jiao Tong University, Shanghai, China(上海交通大学集成电路学院) ; College of Future Information Technology, Fudan University, Shanghai, China(复旦大学未来信息科技学院) ; Aerospace Laser Technology and System Department, Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai, China(中国科学院上海光学精密机械研究所航天激光技术与系统部) ; Fujian Ocean Innovation Center, Xiamen, China(福建海洋创新中心(厦门))
AI总结 本文提出SISTA-Net,通过融合局部和全局特征建模,提升单像素成像的鲁棒性和抗干扰能力,在多个仿真场景中实现2.6dB的PSNR提升,远场水下测试平均PSNR提高3.4dB。
Comments 10 pages, 9 figures, 2 algorithms, 2 tables, journal paper