PINet Lightweight Lane Detection
An innovative approach to lane detection using Point Instance Network (PINet), clustering key points to accurately represent traffic lanes.
An innovative approach to lane detection using Point Instance Network (PINet), clustering key points to accurately represent traffic lanes.
A robust tool for calculating crash data metrics, providing quick insights and valuable analytics.
Exploring semantic segmentation techniques for road scenes using advanced DNN architectures, including GCN, PSPNet, and DUC-HDC.
Top 5% in Kaggle’s challenge to classify icebergs and ships in satellite images with a lightweight model.
Focused on parametric model R&D and data consultancy to enhance supply chain management and demographic analysis.
Developed educational games to teach computational thinking through gamification, enhancing K-12 learning in mathematics and computer science.