End-to-End Optimization Approach for Super-Resolution Imaging of Human Silhouettes through Coded Illumination and Reconstruction Network
Shunsuke Sakoda, Tomoya Nakamura, Yasushi Yagi · 2024
We proposed a super-resolution human-silhouette imaging system using end-to-end optimization of coded illumination and a reconstruction network, increasing resolution without altering camera hardware. We experimentally verified the performance and analyzed it using pseudo-color-based visualization.