Subhasish Basak
Research & Project experiences : 2020
2020 Projects
Find more projects on my GitHub Page!
Find more on GitHubImage Processing Projects
Spring 2020
As a part of my postgraduate coursework at CMI, Computer Vision : CGE-136 I have done some interesting image processing projects. I list some of them as follows :
Image Filtering
Image Pyramids and Image Mosaics
Feature Detection and Description
Object Recognition with Bag of Words
Image Classification with Convolutional Neural Nets
Image Classification using Transfer Learning
Platform/packages used: python, OpenCV, PIL, tensorflow, keras, matplotlib, Numpy and more
Natural Language Processing Projects
Spring 2020
As a part of my postgraduate coursework at CMI, Natural Language Processing : CGE-096 I have done some interesting text analytics projects. I list them as follows :
Building Word Vectors using HAL
Building Semantic Word Vector Models
Building n-gram language model
Word to numeric translation
Recognizing English words with NN
Building Probabilistic language model
Sentiment analysis with RNN & LSTM
Platform/packages used: python, tensorflow, keras, NLTK and more
My Quarantine Projects
Sentiment analysis with WhatsApp emojis
Summer 2020
During the great quarantine of 2020 I started a working on this project which involved several applications of Image processing & sentiment analysis. I developed a command line based python module on GitHub called EmojiScore v0.1.0, which computes the sentiment score of the user upon feeding the screenshot of the recent used emoji box of WhatsApp.
Keywords/Technologies Used: Image processing, Sentiment Analysis, python, vim, Jupyter-notebook, git, GitHub
Designing my New webpage
Summer 2020
The great quarantine of 2020 gave me the time and opportunity to start reconstructing my personal webpage on GitHub pages. I improved my working knowledge on HTML, CSS and Bootstrap. Using several templates available on GitHub I recreated a my webpage (the one you are currently browsing on), it contains detailed information about my work and project experiences and I think it looks more professional than the previous one !
Technologies Used:HTML, CSS, Bootstrap, vim
WhatsApp Chat bot with NLP
Summer 2020
Another fun ML project during 2020 quarantine. We developed a command line based python module on GitHub called ChatBot v0.1.0. Basically a whatsapp bot which can schedule text messages, give automatic reply, send recursive mesages etc. It uses an NLP model based on word-2-vec and empirical probabilistic predictive model.
Keywords/Technologies Used: NLP, python, vim, Jupyter-notebook, git, GitHub
Spotle Ai-thon
Fall 2020
A global challenge to recognize the best talents in Artificial Intelligence and Technology, in association with CMI, SPJIMR and NASSCOM Community. The hackathon is supported by LTI and EKO Informatics. Our team secured global rank 4 & 8, for the two consecutive rounds respectively.
Keywords/Technologies Used: Image processing/CV, tensorflow, python, vim, Jupyter-notebook, git, GitHub
GP MLE Project
L2S, École CentraleSupelec, Université Paris-Saclay
October 2019 - October 2020
Worked on Numerical issues in maximum likelihood parameter estimation for Gaussian process regression with Sebastien Petit (Safran Aircraft Engines) under the guidance of Prof. Julien Bect and Prof. Emmanuel Vazquez (L2S, CentraleSupelec).
Technologies/Keywords Used: GP Regression, MLE, optimization, GPy, python