Pavan Kumar Sathya Venkatesh

I am a First year MSCS student at University of Massachusetts, Amherst. I did my Bachelors in Computer Science and Engineering at VIT University, Chennai, advised by Dr. Pandiyaraju V.

My research focuses on computer vision and medical imaging, where I develop deep learning methods for clinical diagnosis and decision support. I'm particularly interested in uncertainty quantification, multi-modal learning, and neural surface reconstruction.

Previously I interned as a backend developer at Prodapt Solutions Private Ltd. In the summer of 2024, I was an intern at MedxAI mentored by Dr. Susan Elias and Dr. Sheena Pravin.

Email  /  CV  /  Google Scholar  /  GitHub  /  LinkedIn

Highlights

Endobuddy
AI assisted UGI endoscopy

CM-TGT
Multi Modal Graph Based Deception Detection

SPROUT
Few-Shot learning based plant disease identification

CerviLens
Vision based Cervical Cancer Identification

Selected Publications
UGPL: Uncertainty-Guided Progressive Learning for Evidence-Based Classification in Computed Tomography
Shravan Venkatraman, Pavan Kumar S, Rakesh Raj, Chandrakala S,
Accepted (Poster): CVAMD @ ICCV, 2025
project page

A progressive learning framework that leverages uncertainty quantification to guide the network's attention from global context to diagnostically ambiguous regions, enabling precise evidence-based classification in CT scans through adaptive feature refinement and multi-scale analysis.

SPROUT: Symptom-centric Prototypical Representation Optimization and Uncertainty-aware Tuning for Few-Shot Precision Agriculture
Shravan Venkatraman, S Pavan Kumar, Pandiyaraju V, A Abeshek, S A Aravintakshan, Kannan A
Accepted: NeuroComputing, 2025
paper

A few-shot learning approach for plant disease detection that learns symptom-centric prototypical representations through uncertainty-aware optimization, enabling accurate disease classification from limited labeled examples in resource-constrained agricultural settings.

Making Lies Visible: CM-TGT for Graph-Based Cross-Modal Deception Detection
S Pavan Kumar, Maheswer Sunil Kumar, Pranay Jiljith T, Pandiyaraju V
In Progress
code / paper / project page

A cross-modal deception detection framework that constructs temporal graphs to model audio-visual interactions, capturing subtle discrepancies between verbal and non-verbal cues through graph attention mechanisms to identify deceptive behavior in court trials and game show scenarios.


Yep it's another Jon Barron website.
Last updated November 2024.