Skip to content

Find Anything Anywhere All at Once - One-Shot Instance Segmentation of Objects in a Point Cloud

Notifications You must be signed in to change notification settings

mohith-sakthivel/find-anything

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

67 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Find Anything Anywhere All At Once

In this project, we develop a network that can segment intances of a previously unseen object in a point cloud. The network takes as input the scene point cloud and a point cloud corresponding to the desired object, and predicts which points in the source point cloud belong to instances of the diesired object.

Model Architecture

For more details, refer project poster.

Training

To train the model, run

python -m fa.train --seed 0 --exp_name test_exp --batch_size 16

About

Find Anything Anywhere All at Once - One-Shot Instance Segmentation of Objects in a Point Cloud

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages