The project aims to explore feature extraction methods for high dimensional medical scan data like CT and MRI scans. Medical datasets are often a victim of the "curse of dimensionality". We plan to implement modular pipelines for training and testing various Computer Vision methods to gauge their relevance and practicality in prediction models.
- Computer Vision
- Unsupervised Learning
- Representation Learning
- Image Classification/Segmentation
While there are no strict requirements, familiarity with the following will be useful: jupyter notebooks, pytorch, numpy and pandas. Beginners to Deep Learning are welcome to join our team and learn while contributing to the project.
Results are not everything. This may be a fairly long term project with many learnings along the way. Don't expect immediate results and do be ready to work towards a final product.
Ramanathan Rajaraman Tanmay Devale