Real Image Denoising with Knowledge Distillation for High-Performance Mobile NPUs
面向高性能移动NPU的真实图像去噪知识蒸馏
机构 * Computer Vision Lab, CAIDAS & IFI, University of Würzburg, Germany(计算机视觉实验室,CAIDAS与IFI,乌尔姆大学,德国)
AI总结 提出NPU感知的硬件-算法协同设计方法,通过知识蒸馏将高容量教师网络的知识迁移到轻量级学生网络,在移动NPU上实现高效去噪,PSNR仅损失0.05 dB,推理速度比GPU快3.88倍。
Journal ref Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2026, pp. 3792-3800