Daily Report

  • 7/5/21 - 7/9/21 and 7/12/21 - 7/16/21

    • "Read about some methods for point cloud Segmentation:

      • PointNet
      • PointNet++
      • RandLA-Net
      • RSNet (Recurrent Slice Networks)
      • Some other methods

    • Videos of some of the above approaches (for better understanding)

    • Read about:

      • GitPod
      • Github Actions
      • GitPod Customized Docker container
      • Docker
      • Docker Container Action

    • Read a bit about:

      • NVIDIA GPU Cloud
      • NVIDIA Docker

    • Explore NGC Container library

    • Find suitable containers that can be used for ML and AI of point cloud data (Couldn't find anything that looked promising)

    • Read about the basics:

      • Lidar/Structure from motion data types
      • ...

    • ...

  • 7/19/21
    • GAN Paper
    • Video (for more details)
  • 7/20/21
    • DCGAN (Implementation using PyTorch)
    • In person meeting
  • 7/21/21
    • DCGAN (Implementation)
    • Documentation in Github Pages
  • 7/22/21
    • Ubuntu Setup
    • Pytorch Notebook Dockerfile
  • 7/23/21
    • Searching through different problems that we can work on
    • CS231n (CV Course)
  • 7/26/21
    • Visualization of point clouds with CloudCompare
    • Container Camp
    • CS231n (Computer Vision Course)
  • 7/27/21
    • PointNet
    • Container Camp
  • 7/28/21
    • Docker
    • Container Camp
  • 7/29/21
    • PointNet Implementation (Using PyTorch)
  • 7/30/21
    • PointNet Classification Implementation
    • Meeting with Travis on PhytoOracle Project
    • Deep Learning Course
  • 8/2/21
    • Playing with gpu06 (SSH + Docker + ...)
    • Testing My PyTorch Docker Image
    • Testing docker in gpu06
  • 8/3/21
    • Python Scripts of PointNet Classficiation in gpu06 + github
    • Container Camp
    • Deep Learning Course
  • 8/4/21
    • PointNet Classification + Documentation + Plots + ...
    • Deep Learning Course
  • 8/5/21
    • PointNet Segmentation
  • 8/9/21
    • PointNet Segmentation
    • Documentation
    • Find Good labeling tool
  • 8/10/21
    • PointNet++ Paper
    • Deep Learning Course
  • 8/11/21
    • Finding labeling tool
    • Deep Learning Course
    • DL Notebooks
  • 8/12/2021
    • Deep Learning Course
    • Data preparation scripts
  • 8/13/2021
    • Deep Learning Course
    • Pipeline of annotating data
    • Set up new laptop
  • 8/14/2021
    • Setup ubuntu & ...
    • Workning with supervisely
  • 8/16/21
    • splitting 3D Lettuce dataset to batches for labeling
    • Deep Learnin Course
  • 8/17/21
    • Annotating data with supervisely
    • Scripts for conversion in gihub
    • Deep Learnin Course
  • 8/18/21
    • Rand-LA Net Paper (half)
    • Deep Learning Course
    • Annotate a batch of data
    • Create a tool for visualization of labeled pointclouds
  • 8/19/21
    • Add more options to visualization tool
    • Rand-LA Net Paper (half)
    • Deep Learning Course
  • 8/20/21
    • Rand-LA Net PyTorch Implementation
    • Deep Learning Course
  • 8/23/21
    • Rand-LA Net PyTorch Implementation + Results + Github
    • Annotating PhytoOracle Dataset
  • 8/24/21
    • PointNet++ Imeplementation + Results + Github
    • Annotating PhytoOracle Dataset
  • 8/25/21
    • Annotating PhytoOracle Dataset
    • 3D Lettuce Soil Segmentation (PointNet) + Result
  • 8/26/21
    • PointNet++ and RandLA-Net on Lettuce Soil Segmentation + Results
    • Annotating PhytoOracle Dataset
  • 8/27/21
    • Exploring other methods
    • Annotating PhytoOracle Dataset
  • 8/28/21
    • Annotating PhytoOracle Dataset
  • 8/30/21
    • Autoencoder + pytorch implementation
    • playing with annotator
    • Computer Vision Course
  • 8/31/21
    • Variational Autoencoder + pytorch implementation
    • Annotating with Supervisely
    • Computer Vision Course
  • 9/1/21
    • Read a bit about ConvTranspose
    • Annotating with Supervisely
    • Computer Vision Course
  • 9/2/21
    • Preparing for slides and etc. for presentation
  • 9/7/21
    • Computer Vision Course
    • Docker tut + test
  • 9/8/21
    • Computer Vision Course
    • Annotation
  • 9/9/21
    • KNN for upsampling + multiprocessing
  • 9/10/21
    • Dockerization of the project + Scripts + ...
  • 9/13/21
    • Retrain models with new data
    • Test new models in dockerized env.
    • Image on dockerhub
    • Computer Vision Course
  • 9/14/21
    • GUI for Lettuce/Soil PointClouds using tkinter
    • Dockerization of the GUI
    • Test on different batch
    • Computer Vision Course
  • 9/15/21
    • DGCNN Paper
  • 9/18/21
    • DGCNN Paper
    • DGCNN Implementation on airplanes dataset
  • 9/20/21
    • DGCNN re-Implementation on lettuce dataset
    • Improve DGCNN by modifying it and make it simpler (prevent overfitting)
  • 9/21/21
    • Re-Evaluation of all models on a fixed split of dataset
    • Tunning the training of DGCNN
  • 9/22/21
    • Evaluation of PointNet++, RandLANet and DGCNN on a batch of data
    • Some labeling
    • Computer Vision course
  • 9/23/21
    • Dockerize training of the PhytoOracle dataset
    • Trying to use CUDA (for libraries) in built stage of docker (no completed)
  • 9/24/21
    • Containerization of train + predict - ...
  • 9/25/21
    • Computer Vision Course
    • Computer Vision HW (HW1 Q1 Prac.)
  • 9/26/21
    • Computer Vision Course
    • Test Containerized Environment
    • Test Predic on a different batch
    • (Next : Start writing paper)