3D Object Classification using Mesh Neural Network

  • Machine Learning

Mentors :

  • Sudhir Shinde

Mentees :

  • Reet Mehta
  • Yogesh Supe
  • Prathmesh Bele
  • Shalabh Gupta

Implementing mesh neural network for 3D shape representation.

The project will be an implementation of the paper MeshNet: Mesh Neural Network for 3D Shape Representation (AAAI 2019). Mesh is an important and powerful type of data for 3D shapes and widely studied in the field of computer vision and computer graphics. A lot of research is going on to improve the accuracy of the 3D model classification. This paper has achieved 91.9% accuracy. You are expected to implement this research paper. The additional work may include modifying the NN Architecture and train the model on additional datasets.

The following points must be included in the proposal for the project:

- Your motivation and understanding of the project
- Background in ML/DL (include your previous projects)
- Experience with Python and libraries (Numpy, Pandas, TensorFlow, PyTorch, etc)

Tentative Timeline :

Week Work
Week 1 Brush-up Python, ML & Various Deep Learning Architecture
Week 2 Understand the main paper and study input data structure of the 3D model
Week 3 Pre-processing of the data
Week 4 Test the input data on simple DNN model
Week 5 Implement the MeshNet Model
Week 6 Hyperparameter Tuning of the Model
Week 7 Improve the architecture for better performance
Week 8 Generate additional dataset and testing the robustness of the model