search

Diananini / LD-ConGR-CVPR2022

A Large RGB-D Video Dataset for Long-Distance Continuous Gesture Recognition
23 stars 2 forks source link

Usage of depth data #4

Closed erwangccc closed 2 years ago

erwangccc commented 2 years ago

Sorry to bother you again.

The depth data is saved as black/white video. And you load depth and RGB data as images to train and test, right? If so, we loss depth info, do i understand correctly？

Diananini commented 2 years ago

The depth data is normalized to [0, 255].