From Knowledge based Vision Systems to Cognitive Vision Systems: A Review

Computer vision research and applications have their origins in 1960s. Limitations in computational resources inherent of that time, among other reasons, caused research to move away from artificial intelligence and generic recognition goals to accomplish simple tasks for constrained scenarios. In the past decades, the development in machine learning techniques has contributed to noteworthy progress in vision systems. However, most applications rely on purely bottom-up approaches that require large amounts of training data and are not able to generalize well for novel data. In this work, we survey knowledge associated to Computer Vision Systems developed in the last ten years. It is seen that the use of explicit knowledge has contributed to improve several computer vision tasks. The integration of explicit knowledge with image data enables the development of applications that operate on a joint bottom-up and top-down approach to visual learning, analogous to human vision. Knowledge associated to vision systems is shown to have less dependency on data, increased accuracy, and robustness.