Minesh Mathew

I am a ML scientist at Wadhwani AI, where I look into the application of computer vision and document image analysis to problems in public health space. I have Completed MS + PhD from IIIT Hyderabad. I had obtained my undergraduate degree in Btech Computer Science Engineering from NIT Warangal.

My PhD thesis deals with machine-understanding of document images. Specifically, I worked on problems such as OCR in Indian languages, Scene text understanding and Document Visual Question Answering (DocVQA)

News /Updates

  • [May 2023] - Received CVPR 2023 , outstading reviewer award
  • [April 2023] - ICDAR 2023 Challenge on Road text detection , tracking and recognition comes to an end. Results are public now
  • [April 2023] - George’s work on VQA for driving videos accepted at ICDAR 2023
  • [Jan 2023] We are organizing two challenges in ICDAR 2023
    • ICDAR 2023 Competition on RoadText Video Text Detection, Tracking and Recognition
    • ICDAR 2023 Competition on Text-based Video Question Answering on News Videos
  • [Sept 22] - Soumya’s work on VQA on accepted to WACV 2023
  • [May 2022] - Our work comparing different CTC based architectures for Indian languages OCR, now available in arXiv - Link
  • [March 2022] - Our paper “Read while you Drive - Multilingual Text Tracking on the Road” accepted to DAS 2022, Congrats Sergi and George
  • [Oct 2021] Attended ICCV 2021 Doctoral Consortium
  • [Oct 2021] InfographicVQA paper accepted at WACV 2022
  • [Sept 2021] - Presented our work on QA over handwritten documents (oral) at ICDAR 2021
  • [Sept 2021] - Organized first Edition of DocVQA workshop at ICDAR

Academic Services

Achievements / Recognitions

  • [2021] - Outstanding reviewer (top 5%)- ICCV 2021
  • [2021] - Selected for WACV 2021 and ICCV 2021 doctoral consortiums
  • [2017] - Best Paper; Int’l Workshop on Arabic Script Analysis and Recognition (ASAR)
  • [2015] - Runner up - Micorosoft Azure ML Hackathon held at IIIT Hyderabad
  • [2015] - TCS PhD fellowship