Sourajit Saha

PhD Student at University of Maryland, Baltimore County

Email: ssaha2@umbc.edu

Computer Vision | Machine Learning | Robust Visual Perception | Scene Understanding | Data Augmentation | Domain Generalization

Bio

I am currently pursuing PhD in Computer Science under the supervision of Professor Tejas Gokhale at University of Maryland, Baltimore County (UMBC). Prior to that I was a Research Assistant at Center for Cognitive Skill Enhancement (CCSE) at Independent University Bangladesh (IUB). I have received a Masters of Science in Computer Science from University of Maryland, Baltimore County in May 2023 and a Bachelor of Science in Computer Science from BRAC University in August 2017. CV

My primary area of research is Computer Vision and Machine Learning. I am currently studying methods to impose strong shift invariance in CNNs, and mechanisms to circumvent class imbalance in semantic segmentation models. One of my other major area of interest is domain distribution shift for dense prediction (detection, segmentation, depth estimation) models, having methods that circumvent the lack of diversity in academically available datasets on which autonomous vehicle vision systems are trained is a cry for help! I am broadly interested in investigating attention based and knowledge infused vision models in the application of autonomous driving, and other sensory perception tasks. Another recent yet major research interest of mine is Vision + Language. I plan to explore mechanisms for infusing knowledge and visual grounding to build computer vision systems that are robust, efficacious and can interact with human in meaningful ways.

News
  • Feb 2024: Guest lecture on CMSC 491/691: Computer Vision at UMBC.
  • Feb 2024: Reviewing manuscripts for IJCAI 2024, ACM Transactions on Computing for Healthcare.
  • Nov 2023: Guest lecture on CMSC 678: (Graduate) Machine Learning at UMBC.
  • Oct 2023: Reviewing manuscripts for AAAI 2024 (student abstract).
  • Sep 2023: Joined UMBC Robust Perception Group (RPG), lead by Tejas Gokhale.
  • Aug 2023: One paper accepted at ICCV 2023 in OODCV, Paris, France.
  • May 2023: I got my Masters in Computer Science at UMBC in Spring 2023.
  • Oct 2022: One paper accepted at AAAI 2023 in Washington D.C.
  • Mar 2022: One paper accepted at IEEE ACCESS.
  • Oct 2021: One paper accepted at SPIE Medical Imaging Conference 2022 in San Diego.
  • Summer 2021: I was a Machine Learning intern at VIPAR Lab (Vision Image Processing Algorithms Research Lab), UMBC.
  • Jan 2021: I joined UMBC for grad school.
  • Sep 2020: One book chapter published in Part of the Advances in Intelligent Systems and Computing book series, Springer.
  • Dec 2019: One paper accepted at ICMLA 2019 in Florida.
  • Feb 2019: One paper accepted at ICCIT 2018 in Dhaka.
  • Dec 2018: One paper accepted at ELSEVIER - Procedia Computer Science.
  • Jul 2018: I joined IUB as a Research Assistant.
Publications

Most recent publications on Google Scholar.

Improving Shift Invariance in Convolutional Neural Networks with Translation Invariant Polyphase Sampling. Sourajit Saha, Tejas Gokhale

OOD-CV @ ICCV 2023


RFC-Net: Learning High Resolution Global Features for Medical Image Segmentation on a Computational Budget (Student Abstract). Sourajit Saha, Shaswati Saha, Md Osman Gani, Tim Oates, David Chapman

AAAI 2023 Paper Code


Mitigating Domain Shift in AI-Based TB Screening With Unsupervised Domain Adaptation. Nishanjan Ravin, Sourajit Saha, Alan Schweitzer, Ameena Elahi, Farouk Dako, Daniel Mollura, David Chapman

IEEE Access Paper Code


Pairwise Meta Learning Pipeline: Classifying COVID-19 abnormalities on chest radio-graphs. Sourajit Saha, Yaacov Yesha, Yelena Yesha, Aryya Gangopadhyay, David Chapman, Michael Morris, Babak Saboury, Phuong Nguyen

SPIE Medical Imaging Conference 2022 Paper


A comprehensive set of novel residual blocks for deep learning architectures for diagnosis of retinal diseases from optical coherence tomography images. Sharif Amit Kamran, Sourajit Saha, Ali Shihab Sabbir, Alireza Tavakkoli

Springer Book Series, 2020 Paper Code


Optic-Net: A Novel Convolutional Neural Network for Diagnosis of Retinal Diseases from Optical Tomography Images. Sharif Amit Kamran, Sourajit Saha, Ali Shihab Sabbir, Alireza Tavakkoli

ICMLA 2019 Paper Code


A Lightning fast approach to classify Bangla Handwritten Characters and Numerals using newly structured Deep Neural Network. Sourajit Saha, Nisha Saha

Procedia Computer Science, ELSEVIER Paper


Total recall: understanding traffic signs using deep convolutional neural network. Sourajit Saha, Sharif Amit Kamran, Ali Shihab Sabbir

ICCIT 2018 Paper Code

Academic Service

    Membership

  • IEEE, AAAI, ACM, SPIE
Acknowledgement

Template borrowed from Martin Saveski.