search

bradenhurl / DeepGTAV-PreSIL

GNU General Public License v3.0
76 stars 16 forks source link

Could this simulator get 3D semantic point clouds like Semantic-KITTI dataset? #13

Open kxhit opened 4 years ago

kxhit commented 4 years ago

Hi! Thanks for sharing the codes! Semantic segmentation task on 3D point clouds is attracting more and more researchers nowadays. However, the annotation of real world point clouds is expensive. The Semantic-KITTI dataset is the only one with semantic point cloud sequences. I'm curious of if this tool could get the a synthetic dataset like semantic-kitti. Thanks for giving hints to me! Discussions are welcomed!

bradenhurl commented 4 years ago

My lab also created Unlimited Road-scene Synthetic Annotation (URSA) Dataset. I think this is similar to what you're looking for, check it out here: https://uwaterloo.ca/waterloo-intelligent-systems-engineering-lab/ursa

kxhit commented 4 years ago

Thanks for your response! There appears to be no lidar point clouds in this Dataset.

zeng-hello-world commented 3 years ago

Similiar question in point cloud semantic segmentation.

bradenhurl commented 3 years ago

This repo can determine the points belonging to specific objects. Past that detail it would require significant work to be merged with a semantic segmentation method for GTAV. If you merge this repo with a method which does semantic segmentation for images, then you will be able to easily obtain semantic segmentation for the pointclouds since the PreSIL LiDAR is generated from the depth map (thus you will already have the image pixel for the 3D point).