why not directly use the original image pixel values as the r, g, and b values of generated virtual points?

hailanyi / VirConv

Virtual Sparse Convolution for Multimodal 3D Object Detection

https://arxiv.org/abs/2303.02314

Apache License 2.0

276 stars 39 forks source link

why not directly use the original image pixel values as the r, g, and b values of generated virtual points? #43

Open Slonna opened 1 year ago

Slonna commented 1 year ago

Hi ,congratulations on doing such an excellent job! I need your help with a question,why not directly use the original image pixel values as the r, g, and b values of generated virtual points using the completed depth, but actually instead use the results after divided by 3 the confusing code