• 空間特性を活用した手法 “SAMC” [1] • 深層学習に基づく手法 “MTTS-CAN” [12] • 時空間特性を階層的に活用した手法 “Hierarchical” [13] 17 データセットの詳細 データセット [9] Maki+, Inter-beat interval estimation from facial video based on reliability of bvp signals, EMBC 2019 [10] Nowara+, Sparseppg: Towards driver monitoring using camera-based vital signs estimation in near-infrared, CVPRW 2018 [11] Bobbia+, Unsupervised skin tissue segmentation for remote photoplethysmography, Pattern Recognition Letters 2019 比較手法 [1] Kurihara+, Blood Volume Pulse Signal Extraction based on Spatio-Temporal Low-Rank Approximation for Heart Rate Estimation, VCIP 2022 [12]Liu+, Multi-task temporal shift attention networks for on-device contactless vitals measurement, NeurIPS 2020 [13]Tulyakov+, Self-adaptive matrix completion for heart rate estimation from face videos under realistic conditions, CVPR 2016 Tokyo [9] MR [10] UBFC [11] 被験者数 9 8 47 映像数 9 8 50 解像度 640×480 640×640 640×480 フレームレート 30 fps 30 fps 30 fps 映像の長さ 180 s 180 s 60 s UBFC #5 Tokyo #3 MR #1 撮影例