Learning-Based Robot Vision

Focusing on the development of autonomous camera-equipped robots, the book outlines a practical methodology for enabling these systems to perform specific tasks in open environments. It emphasizes the importance of training, interaction, and real-world learning to acquire necessary skills. The author supports the methodology with a structured case study that includes key sub-tasks like localizing, approaching, grasping, and carrying objects, demonstrating the effectiveness of visual demonstration and neural learning in achieving situated competences.