Thanh Le

University of North Texas, University of California, San Francisco

Papers

Total Citations

H-Index

About

Thanh Le is a computer vision researcher whose work addresses fundamental challenges in visual tracking and image feature extraction. His most significant contribution is the introduction of **PlanarTrack**, a large-scale benchmark dataset for planar object tracking published in 2023. This work addresses a critical gap in the field—the lack of challenging, standardized datasets for evaluating deep learning-based planar trackers, which are essential for applications in robotics and augmented reality. Although recently published, PlanarTrack has already garnered 3 citations, signaling its growing importance as a reference resource. Earlier, Le developed a **probability-based approach for multi-scale image feature extraction** (2014, 2 citations), tackling the persistent problem of accurate shape boundary localization. This work improved upon traditional deformable active contour (snake) methods, offering a more robust solution for object tracking, content-based retrieval, and biomedical imaging. By bridging classical feature extraction techniques with modern deep learning benchmarks, Le’s research provides foundational tools for advancing visual perception systems. His work is particularly valuable for students and researchers seeking to understand both the historical evolution and current frontiers of planar object tracking and feature extraction.