《A Reinforcement Learning Framework for Medical Image Segmentation》.pdf

2006 International Joint Conference on Neural Networks Sheraton Vancouver Wall Centre Hotel, Vancouver, BC, Canada July 16-21, 2006 A Reinforcement Learning Framework for Medical Image Segmentation Farhang Sahba, Member, IEEE, and Hamid R. Tizhoosh, and Magdy M.A. Salama, Fellow, IEEE Abstract— This paper introduces a new method to medical Considering the above factors our new algorithm based on image segmentation using a reinforcement learning scheme. reinforcement learning (RL) is introduced to locally segment We use this novel idea as an effective way to optimally find the prostate in ultrasound images. The most important con- the appropriate local thresholding and structuring element values and segment the prostate in ultrasound images. Re- cept of RL is learning by trial and error based on interaction inforcement learning agent uses an ultrasound image and with the environment [3], [4]. It makes the RL agent suitable its manually segmented version and takes some actions (i.e., for dynamic environments. Its goal is to find out an action different thresholding and structuring element values) to change policy that controls the behavior of the dynamic process, the environment (the quality of segmented image). The agent guided by signals (reinforcements) that indicate how well it is provided with a scalar reinforcement signal determined objectively. The agent uses these objective reward/punishment has been performing the required task. to explore/exploit the solution space. The values obtained using In the case of applying this method to medical image this way can be used as valuable knowledge to fill a Q-matrix. segmentation, the