FeverNet: Enabling accurate and robust remote fever screening

Mengkai Yan, Jianjun Qian, Hang Shao, Lei Luo, Jian Yang

Abstract: Highlights•We propose a novel fever screening framework FeverNet, which enables accurate and robust fever screening, without using any external sensors.•We develop the depth estimation network (DEN) to prevent measurement errors caused by distance changes.•We propose the Temperature Distribution Constraints across Time (TDCT) to enhance environmental adaptability.•We propose the Fever Attention Module (FAM) to enhance feature representation and expand the contrastive difference between fever and normal faces.