search

weiyithu / SurroundDepth

[CoRL 2022] SurroundDepth: Entangling Surrounding Views for Self-Supervised Multi-Camera Depth Estimation
MIT License
258 stars 38 forks source link

Why the number of training datasets is less than the number of official nuScene dataset? #7

Open rockywind opened 2 years ago

rockywind commented 2 years ago

Hi, I found the number of training datasets is 20096, but the number of offical nuScene datasets is 28119. I am confused why the training number is less than the official nuScene dataset. And another question is how to generate the train.txt? Thank you for your help.

weiyithu commented 2 years ago

We filter out static frames in nuScenes dataset. Specifically, if the scale of translation between current frame and previous frame or current frame and next frame is smaller than 0.1m, we will delete this frame.

rockywind commented 2 years ago

Thank you very much!