Kangru Wang

Papers

Total Citations

H-Index

About

Kangru Wang is a rising researcher in computer vision, with a primary focus on 3D object detection and its application in dynamic, real-world environments such as autonomous driving and robotics. His most notable contribution, the paper "I3DOD: Towards Incremental 3D Object Detection via Prompting" (2023), tackles the critical challenge of catastrophic forgetting in class-incremental learning scenarios. While traditional 3D detection models excel in static settings, they fail when asked to recognize new object classes without forgetting previously learned ones. Wang’s work introduces a prompting-based framework that allows models to adapt incrementally, preserving performance on old classes while learning new ones—a breakthrough for systems that must evolve over time. Although early in its impact, with 2 citations, this work addresses a fundamental gap in continual learning for perception systems. Wang’s research sits at the intersection of lifelong learning and 3D vision, promising to make autonomous systems more robust and adaptable. His contributions are particularly relevant for students and engineers working on scalable, real-time perception in robotics and self-driving cars.