Application of Opposition-Based Reinforcement Learning in Image Segmentation
Farhang Sahba, Hamid R. Tizhoosh, Magdy M Salama · 2007
In this paper a method for image segmentation using an opposition-based reinforcement learning scheme is introduced. We use this agent-based approach to optimally find the appropriate local values and segment the object. The agent uses an image and its manually segmented version and takes some actions to change the environment (the quality of segmented image). The agent is provided with a scalar reinforcement signal as reward/punishment. The agent uses this information to explore/exploit the solution space. The values obtained can be used as valuable knowledge to fill the Q-matrix. The results demonstrate potential for applying this new method in the field of medical image segmentation