Computer Vision & Deep Learning Projects

Bayesian Active Learning for Semantic Segmentation

arXiv link

Introduces Beta distribution approximation for efficient uncertainty estimation in semantic segmentation, achieving significant annotation cost reductions through intelligent sample selection.

Self-Supervised Contrastive Learning for 3D Mesh Segmentation

arXiv link

Pioneering self-supervised approach for 3D geometric data analysis, eliminating the need for manual annotations in mesh segmentation tasks through contrastive learning techniques.

Active Learning for Medical Imaging

Published in Frontiers in Radiology, 2021

Clinical validation demonstrating dramatic reduction in radiologist annotation burden while maintaining diagnostic accuracy through uncertainty-guided active learning frameworks.

Multi-View 3D Semantic Segmentation

Probabilistic framework for 3D scene understanding that leverages multiple viewpoints to improve segmentation accuracy while quantifying prediction uncertainty.

Uncertainty-Aware Computer Vision

Development of computer vision systems that provide reliable confidence estimates for critical applications in medical imaging and autonomous systems.

Real-Time Defect Detection

Industrial computer vision systems for manufacturing quality control, including 3D CT composite data analysis and automated defect recognition.