Affordance Segmentation

Affordance segmentation aims to identify image regions indicating how an object can be interacted with, crucial for robotics and human-computer interaction. Current research focuses on improving the accuracy and robustness of segmentation models, particularly addressing challenges like scale variations, hand occlusions, and the need for less annotation, often leveraging architectures like Mask-RCNN and Vision Transformers, along with the incorporation of large vision-language models. This work is significant because accurate affordance segmentation enables more intuitive and effective human-robot interaction, leading to advancements in robotic manipulation and assistive technologies.