News

⭐️ [Sep 2019] !NEW! ♨️ ♨️ Started my Research Internship in Natural Language Processing at Snap Inc., Los Angeles, CA

⭐️ [May 2019] !NEW! ♨️ ♨️ Started my summer internship in Natural Language Processing R & D group at Siemens, Princeton, NJ

⭐️ [Aug 2018] ♨️ ♨️ Received the ORC Doctoral Fellowship from the Office of Research and Commercialization

⭐️ [May 2018] ♨️ Two Research Poster Papers got accepted to the 40th International Engineering in Medicine and Biology Conference (EMBC '18)

⭐️ [Mar 2018] ♨️ Secured 11th position among 425 registered participants in the Breast Cancer Histology Classification (ICIAR 2018) organized by AIMI

⭐️ [Mar 2018] Submitted the results for Diabetic Retinopathy Segmentation and Grading Challenge (IDRID 2018) with 0.93 DICE Index on Optical Disc segmentation and 87.5 % accuracy on Diabetic Macular Edema (DME) grading in ISBI 2018

⭐️ [Dec 2017] ♨️ Secured 2nd position in the Auto Rickshaw Detection Challenge, organized by NCVPRIPG ‘17

⭐️ [Sep 2017] ♨️ Secured 8th position out of 4000 registered participants in the "Identifying the product category (Deep Learning) Challenge Sep-2017" organized by Hackerearth

⭐️ [Sep 2017] Submitted the results for Polyp Detection in Endoscopy Images challenge with 87.31% validation accuracy and sensitivty of 92.05% in MICCAI 2017

⭐️ [Apr 2017] Submitted the results for Lung Cancer Detection challenge in Kaggle Data Science Bowl 2017

Recent Publications

Search Publications

Projects

Hate Speech Classification

One of the key challenges in internet content management is the filtering of the profane language. What forms the profane language is the question we got to answer. To do this we categorize it into classes such as offensive language, hate speech, and neither. This categorization is done by...

Diabetic Retinopathy Segmentation and Grading Challenge (IDRID 2018)

Diabetic Retinopathy is the most prevalent cause of avoidable vision impairment, mainly affecting working age population in the world. The aim of this challenge is to evaluate algorithms for automated detection and grading of diabetic retinopathy and diabetic macular edema using retinal fundus images.

Breast Cancer Histology Classification (ICIAR 2018)

Breast cancer is one of the leading cancer-related death causes worldwide, specially woman. However, early diagnosis significantly increases treatment success. The challenge is to automatically classifying H&E stained breast histology microscopy images in four classes: normal, benign, in situ carcinoma, and invasive carcinoma.

Autorikshaw Detection Challenge (NCVPRIPG-2017)

As the autonomous navigation is getting closer to reality and in order address, the safety issues in driving an autorickshaw on Indian roads, the first step would be the high accuracy detection of autorickshaws having a variety of shapes and sizes found on Indian roads, within an image.

Mouse Brain Architecture Project

This is a high-throughput experimental project, the goal of which is to systematically do injection-based tractography in whole mouse brains, on a grid of injections that cover the whole brain.

Identifying the product category (Hackerearth Deep Learning Challenge Sep-2017)

In this challenge, one of the largest retailers in Germany wants to improve their inventory-management process in its Food and Groceries business. The company is looking for intelligent solutions that can reduce the amount of human effort in its warehouse and retail outlets.

Polyp detection in Endoscopy images (MICCAI 2017)

The goal of this project is to produce a system that takes colonoscopy images as input and detects the presence/absence of polyps in each frame using Convolutional Neural Networks (CNN) so as to be used in the prevention of colorectal cancer clinically.

Lung cancer detection (Kaggle Data Science Bowl 2017)

The goal was to find the likelihood of a given 3D CT image having Lung cancer.

EPR and NMR/DNP Three dimensional Spectral-Spatial Image processing and reconstruction

The objective of the project is to develop applications to process and reconstruct three-dimensional Electron Paramagnetic Resonance (EPR), Nuclear Magnetic Resonance (NMR) / Dynamic Nuclear Polarization (DNP) datasets.

Imagined Speech Classification using EEG

The objective of this work is to assess the possibility of using (Electroencephalogram) EEG for communication between different subjects.

Assistive Device for the Speech Impaired

Speech impaired people have difficulty in communicating with normal people because the hand gestures used by them to communicate their information is not easily understandable; only trained people can understand these. Most expressions and emotions remain unconveyed, sometimes even misinterpreted.

Affiliations

Research & Development groups

  • Healthcare Technology Innnovation Centre, Indian Institue of Technology, Madras
  • MRI-MRS Centre, Indian Institue of Technology, Madras
  • Photonics research group, Indian Institue of Technology, Madras

Industrial Organizations

Universities

Contact