Annan Li

About

Annan Li is a leading researcher in computer vision, with a focus on predictive scene parsing and temporal action understanding in video. His work addresses critical challenges in vision-based AI systems, particularly for autonomous driving and robotics. Li’s highly cited paper, “STC-GAN: Spatio-Temporally Coupled Generative Adversarial Networks for Predictive Scene Parsing” (2020, 52 citations), introduced a novel framework that jointly models spatial and temporal dynamics to assign pixel-level semantic labels to future video frames—a task essential for anticipating scene changes in dynamic environments. This contribution significantly advanced the field of predictive scene understanding. In parallel, Li’s research on “Atrous Temporal Convolutional Network for Video Action Segmentation” (2019, 13 citations) tackles fine-grained action segmentation in untrimmed videos, proposing a robust architecture that handles varying temporal scales of human actions. His work has practical implications for surveillance and robotics, where accurate, real-time action recognition is vital. Through these innovations, Li has established himself as a key contributor to spatio-temporal modeling, pushing the boundaries of how machines perceive and predict visual information over time.

Research Focus

Key Achievements

Top Papers

1. 1
   STC-GAN: Spatio-Temporally Coupled Generative Adversarial Networks for Predictive Scene Parsing

   52 citations · 2020
2. 2
   Atrous Temporal Convolutional Network for Video Action Segmentation

   13 citations · 2019

Key Collaborators

Contact & Links