Skip to content

Project for Computational Aspects of Robotics Course from Columbia University's School of Engineering and Applied Science, March 2023

Notifications You must be signed in to change notification settings

alicediakova/NN-RGB-Image-Segmentation

Repository files navigation

NN-RGB-Image-Segmentation

Project for Computational Aspects of Robotics Course from Columbia University's School of Engineering and Applied Science, March 2023

In this homework, I trained a neural network to perform segmentation using RGB images and corresponding ground truth instance segmentation masks obtained from a simulated camera. I then used the model and RGB-D images to estimate the pose of objects.

Pytorch (https://pytorch.org/docs/stable/nn.html) Tutorials Used:

Project Parts:

  1. Generate training set (function compute_camera_matrix in camera.py)
  2. CNN for Segmentation (dataset.py, model.py, segmentation.py)
  3. Pose Estimation (icp.py)

About

Project for Computational Aspects of Robotics Course from Columbia University's School of Engineering and Applied Science, March 2023

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages