International Display Workshops General Incorporated Association

[AIS7/VHF6-3L] Gesture Classification of Single-Pixel-Imaging Reconstruction by Using Deep Learning

AIS Topical Session

*Hiroki Takatsuka1 (1.Utsunomiya University (Japan))

single-pixel imaging, deep learning, gesture classification

https://doi.org/10.36463/idw.2021.1053

We report classification of images reconstructed with different number of masks using single pixel imaging. The target images to be classified are hand gestures (rock, scissors, paper). Deep learning based on LeNet is used for classification. Images that were reconstructed with single-pixel imaging by use of with 400 or more masks were accurately classified by 90% or more.