projects | Tushar Nayak

Neural Active Contours

alpha Medical Imaging Segmentation Active Contours

A non U-Net segmentation approach for medical images

Deformable 3D Surgical Scene Reconstruction

active Surgical Robotics 3D Vision Gaussian Splatting

A surgical vision prototype that combines depth-based point clouds, deformable 3D Gaussian splatting, and VLM-guided semantic querying for endoscopic scenes.

Open-Vocabulary Surgical Tool Detection and Tracking

active prototype Surgical Robotics Tool Tracking Open-Vocabulary AI

A real endoscopy pipeline that turns text prompts into surgical tool boxes, masks, tracking overlays, and evaluation metrics using Grounding DINO and SAM2.

Physics-Informed Endovasculature Deformation Estimation And Registration

ongoing Surgical Robotics Medical Imaging 3D Vision

Research at CERLAB (master's thesis)

Monkeypox Virus Detection Using Skin Lesion Images

published Medical Imaging Dermatology Classification

Prelimnary project at Biomedical Computing Lab

Full-Volume Lung CT Segmentation and Surface Export

completed Medical Imaging CT Segmentation

A real-data MONAI and VTK pipeline that segments full chest CT volumes, exports lung surfaces, and reports validation metrics for navigation-prep experiments.

Skin Lesion Segmentation on ISIC 2018

active extension Medical Imaging Dermatology Segmentation

A reproducible dermoscopy segmentation benchmark with DeepLabV3 baselines, planned U-Net and transformer comparisons, and promptable SAM-style extensions.

Gaussian Occupancy Cardiac Reconstruction

active extension 3D Vision Medical Imaging Cardiac Reconstruction

A sparse-view echocardiography project that recovers 3D cardiac anatomy with Gaussian occupancy fields, differentiable slice supervision, and mesh evaluation.

Sparse-View Coronary Vessel Reconstruction

research prototype 3D Vision Medical Imaging Vascular Reconstruction

A research prototype for recovering 3D coronary vessel geometry from limited angiographic projections using differentiable rendering and hemodynamic constraints.

Spatiotemporal Glioblastoma Evolution Visual Prediction

paused Medical Imaging Neural ODEs MRI

Research at The ∀ Lab & Image Science Lab

Attention-augmented Dual-View Mammogram Alignment for Enhanced BI-RADS Classification

completed Medical Imaging Mammography Classification

Course project for Projects in Biomedical AI

Neural Anisotropic Diffusion for Medical Image Relaxation

completed Medical Imaging PDEs Denoising

Course project for Medical Image Analysis

Neural Morphological Analysis of Fungi

alpha Pathology Microscopy Morphology

A microscopy project for fungal species classification and future neural-field modeling of morphology from high-resolution biological image patches.

Multi-Scale Lung Histopathology Classification

prototype Pathology Histopathology Attention

A pathology model that fuses 20x and 40x whole-slide image views with cross-scale attention for malignancy, subtype, and differentiation grading.

Few-Shot 2D Echo to 3D Cardiac Reconstruction via Neural Implicit Priors

completed 3D Vision Medical Imaging Cardiac Reconstruction

Course Project for Learning for 3D Vision

Multi-Model Oral Squamous Cell Carcinoma Detection

completed Pathology Cancer Detection Deep Learning

Undergraduate thesis

CTA to Mesh & Point-Cloud Coronary Artery Segmentation using Residual 3D U-Net

completed Medical Imaging 3D Vision CT

Course Project for Image Based Computational Modelling & Analysis

Federated Graph Learning for Pathology Images

prototype Pathology Federated Learning Graph Learning

A pathology learning pipeline that combines self-supervised image encoding, patient-level federated training, graph reasoning, and concept-bottleneck classification.

Lung & Colon cancer detection using Histopathological Imaging

published Pathology Histopathology Cancer Detection

Research at Biomedical Computing Lab

Neural Correlates Associated to Images for Emtional Response

completed Medical Imaging fMRI Neuroimaging

Course Project for Fundamentals of MRI and Neuroimaging Analysis