Human Centric Computer Vision

Human-centric computer vision focuses on developing computer vision systems that accurately and robustly understand and interact with humans in images and videos. Current research emphasizes improving model performance in tasks like pose estimation and activity recognition, often leveraging techniques like data augmentation (including synthetic data generation) and incorporating language models to enhance understanding of complex human interactions. These advancements are crucial for applications ranging from human-robot interaction and assistive technologies to improving the fairness and robustness of AI systems used in various societal contexts, particularly by addressing biases and privacy concerns inherent in real-world datasets.