My name is Tanish Mittal. Currently, I am pursuing MSc. Physics and B.E. Computer Science in BITS Pilani, Pilani campus. I am interested in Deep learning and Physics as whole. I also have fair knowledge of data structures and algorithms. My preferred programming languages are Python, C++ and MATLAB.
Here is the list of some DL projects that i have done so far other than tutorials -
- Masked Autoendcoder - This work interpolates the existing Masked Autoencoder ( Published by Facebook research) to the medical domain. This is the first step of the project which has been successful. After this model will be interpolated to 3D images(still in progress).
- TiramisuSeg_Pytorch - Implemented One hundred Layers Tiramisu on COCO dataset using PyTorch for Segmentation
- Removing Background and Cartoonising - This project was a national level competition (HackCV, Roboweek by NIT Hamirpur) in which i was able to secure rank in top 5. In this project, a UNET architecture was implemented for Human semantic segmentation for removing background. Then, the image was cartoonised using 2 GANs and VGG19. For the first part, human images from COCO dataset were used. For second part stills from animes, anime faces dataset from kaggle, 5000 landscape photos and 10000 human face photos from FFHQ dataset were used
- Glitch Classification of signals from LIGO using CNN - This project was my attempt to implement GravitySpy project. I implemented CNN on Gravity SPY dataset for classifiying the glitch in 22 classes. I was able to achieve an accuracy rate of 92%.
- Polynomial Regression in MATLAB - Now, polynomial regression is very basic thing but speciality of this project was that this project do not use any library, complete code was typed by me and it involves 2 methods of gradient descent - stochastic gradient descent, mini batch gradient descent.
- English to Hindi translation - I tried implementing seq2seq learning for translation from English to Hindi