🎉 Completed Internship at Centific.
🎉 Got awarded the ECCV 2026 Travel Grant.
🎉 Paper accepted to ECCV 2026.

Shreya Biswas

Fourth Year PhD Candidate in Computer Science

Few-Shot Computer Vision, Multimodal Foundation Models, 3D Reconstruction and Neural Scene Representation, Data-Efficient Medical Vision

LinkedIn GitHub Email Download CV
Shreya Biswas Profile Photo

About Me

I am a 4th year PhD candidate at Stony Brook University, under the mentorship of Dr. Zhaozheng Yin. My research focusses on few-shot learning, Multimodal Foundation Models, 3D Reconstruction and Neural Scene Representation, Data-Efficient Medical Vision.

Previously, I interned at Centific, where my work included egocentric video simulation dataset generation using Gaussian Splatting, reducing hallucination in video captioning using Cosmos3-Nano Reasoner and Qwen-Instruct vLLM, 3D face pose estimation for selfie videos, and Indian-language OCR for tables and graph understanding.

Education

PhD in Computer Science

SUNY Stony Brook | 2023 - 2028 (Expected)

B.E. in Electronics and Telecommunication

Jadavpur University | 2019 - 2023 | GPA: 9.29/10

Research Areas

Multimodal Foundation Models

Adapting vision-language and foundation models for data-efficient visual understanding.

3D Reconstruction and Neural Scene Representation

Building 3D reconstruction pipelines from real-world egocentric data to generate simulated environments, reconstructed assets, and hand-object interactions.

Data-Efficient Medical Vision

Developing structure-aware AI systems for medical and scientific imaging for medical diagnosis.

Publications

Parkinson's Disease Detection Diagram

Mitigating Pose-Scale Discrepancy Bias and Reforming Multi-Support Reasoning for Few-Shot Semantic Segmentation

ECCV 2026

PDF Webpage Code
Parkinson's Disease Detection Diagram

DOTGraph: CLIP-Driven Feature Disentanglement and Optimal Transport based Graph Learning for Few-Shot Segmentation

WACV 2026

PDF Webpage Code Cite
Parkinson's Disease Detection Diagram

Structure-Aware Vessel-Guided Shadow Removal in 3D OCTA

Under Review, ACCV 2026

Parkinson's Disease Detection Diagram

TaSP: Target Style Guided Pruning for Cross Domain Few Shot Segmentation

Under Review, IEEE TIP

Parkinson's Disease Detection Diagram

Granulated Mask-RCNN and Eye Detection Index (EDI) for detection and localization of the eye of tropical cyclones from satellite imagery

Journal of Data, Information and Management 2024

PDF Webpage Cite
SCENIC Hardware Accelerator Diagram

SCENIC: An Area and Energy-Efficient CNN-based Hardware Accelerator for Discernable Classification of Brain Pathologies using MRI

VLSID 2022

PDF Webpage Cite
Parkinson's Disease Detection Diagram

Moth-flame Optimization based Deep Feature Selection for Cardiovascular disease Detection using ECG Signal

Handbook of Moth-Flame Optimization Algorithm, 2022

PDF Webpage Cite
Parkinson's Disease Detection Diagram

An Ensemble of CNN Models for Parkinson's Disease Detection Using DaTscan Images

Diagnostics, 2022

PDF Webpage Cite
Breast Cancer Detection Diagram

Breast cancer detection from thermal images using a Grunwald-Letnikov-aided Dragonfly algorithm-based deep feature selection method

Computers in Biology and Medicine, 2021

PDF Webpage Cite
COVID-19 CT Detection Diagram

Prediction of COVID-19 from Chest CT Images Using an Ensemble of Deep Learning Models

Applied Sciences, 2021

PDF Webpage Cite
Multi-Level Image Segmentation Diagram

Multi-Level Image Segmentation Using Kapur Entropy Based Dragonfly Algorithm

ISDA 2022

PDF Webpage Cite

Featured Projects

Coastal Flooding Visualizer

Research Project | 2025 - Present

  • Engineered and deployed a geospatial web app (Flask + Leaflet + GEE) to simulate global coastal flooding at 30m resolution.
  • Integrated >10 GB of sea level rise and hurricane datasets for real-time hazard mapping and interactive analytics.
  • Processed 90+ years of hurricane records with <3s query response time for projections and storm overlays.
  • Enabled chatbot-assisted decision support for local-scale climate risk assessment.

WHISPER for Parkinson's

CSE 538 NLP Project | 2024

  • Fine-tuned Whisper on Parkinson's speech data, improving dysarthric speech recognition by 49.1% over baseline.
  • Reduced Word Error Rate by 15%, advancing assistive speech technologies.

Prediction of Video Sequences

Undergraduate Thesis Project | 2023

  • Designed a predictive video-sequence framework using Taylor series, reducing error by 12% vs. baselines.
  • Introduced RL-based Adaptive Probabilistic Learning Matrix, improving stability by 20%.

Research Experience

2023-Present

Graduate Researcher

Stony Brook University (SUNY) Advisor: Dr. Zhaozheng Yin

  • Devised a geometry-aware few-shot segmentation framework by disentangling semantics from instance-level pose.
  • Devised an Optimal Transport-based Graph Neural Network pipeline with CLIP-driven foreground disentanglement and a Graph Contrastive Loss.
  • Designed a topological framework to remove long tailed shadow from 3D OCTA volumes.
  • Developed a structure-preserving style transfer framework with a novel pruning mechanism for cross-domain few-shot classification .
Fall 2023

Undergraduate Research Intern

Indian Statistical Institute (ISI) Advisor: Dr. Sankar K. Pal

  • Proposed integrating granulation in Mask-RCNN to detect cyclone centers (34% increase in detection accuracy, 11% decrease in false alarm rate, 56% increase in segmentation compactness)
Spring 2022

Summer Research Intern

Machine Intelligence Research Labs (MIR Labs) Advisor: Dr. Ajith Abraham

  • Developed a nature-inspired multi-threshold image part-segmentation technique combining Dragonfly Algorithm and Kapur's Entropy.
Fall 2021

Computer Vision Research Intern

SUNY Polytechnic University Advisor: Dr. Janet Paluh

  • Collaborated with 3 neurosurgeons to devise an energy-efficient (reduction by 24\%, 0.265 MB memory space) model for glioma tumor detection (98.3\%) and classification (99.6\%) from MRI modalities.
Fall 2020

Winter Research Intern

Indian Institute of Technology (IIT) Kharagpur Advisor: Dr. Sudipta Mukhopadhyay

  • Implemented deep hashing methods for large-scale image retrieval.

Professional & Leadership Experience

Graduate Teaching Assistant

Department of computer Science 2021 – 2022

  • CSE 532: Theory of Database Systems - under Dr. Fusheng Wang
  • CSE 519: Introduction to Data Science - under Dr. Steven Skienna

Chairperson

IEEE Jadavpur University Computer Society Chapter 2022 – Jan 2023

  • Led AICSSYC’22 with 300+ participants, coordinating speakers, program, and logistics.
  • Organized “Machine Learning Accelerator Summit 2.0”, focusing on hands-on ML workflows.

Secretary

IEEE Jadavpur University Computer Society Chapter 2021 – 2022

  • Produced the “Pass the Mic ’21” podcast series highlighting student research journeys.
  • Co-organized “DoubleSlash” 48-hour hackathon; managed judging rubric and outreach.

Technical Skills

Programming Languages

Python Java C/C++ Matlab HTML/CSS JavaScript NodeJS React Flask

Libraries & Frameworks

PyTorch TensorFlow Keras OpenCV Pandas Numpy Matplolib

Tools & Platforms

VS Code Google Earth Engine Git Google Cloud Platform