期刊
IEEE ACCESS
卷 11, 期 -, 页码 34544-34553出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/ACCESS.2023.3263496
关键词
Measurement; Deep learning; Real-time systems; Computer architecture; Distortion; Convolutional neural networks; Imaging; HDR imaging; objective metrics; no-reference
Efficiency and efficacy are desirable properties for evaluation metrics in SDR and HDR imaging, but it is challenging to satisfy both simultaneously. Existing metrics like HDR-VDP 2.2 accurately mimic the HVS but are computationally expensive, while cheaper alternatives fail to capture crucial aspects of the HVS. In this work, we propose NoR-VDPNet++, a deep learning architecture that converts full-reference metrics to no-reference metrics, reducing computation burden and successfully applied in different scenarios.
Efficiency and efficacy are desirable properties for any evaluation metric having to do with Standard Dynamic Range (SDR) imaging or with High Dynamic Range (HDR) imaging. However, it is a daunting task to satisfy both properties simultaneously. On the one side, existing evaluation metrics like HDR-VDP 2.2 can accurately mimic the Human Visual System (HVS), but this typically comes at a very high computational cost. On the other side, computationally cheaper alternatives (e.g., PSNR, MSE, etc.) fail to capture many crucial aspects of the HVS. In this work, we present NoR-VDPNet++, a deep learning architecture for converting full-reference accurate metrics into no-reference metrics thus reducing the computational burden. We show NoR-VDPNet++ can be successfully employed in different application scenarios.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据