Tissue Segmentation in Medical Images Based on Image Processing Chain Optimization
Shahryar Rahnamayan, Zaid Mohamad · 2010
Segmentation is a crucial task in medical image processing. The accuracy of the segmentation can directly affect other post processing tasks, such as image analysis and feature extraction. Knowledge and sample based learning approaches play a pivotal role in an image processing. However, the acquisition and integration of expert knowledge (for the former) and providing a sufficiently large number of training samples (for the latter) are generally hard to perform and time consuming tasks. Hence, learning image processing tasks from a few gold/ground truth samples (three for the current work), prepared by the radiologist, is highly desirable. The purposed approach utilizes Differential Evolution (DE) to optimize an image processing chain; which has successfully been used to segment breast ultrasound and X-ray lung images. The training is based on three sample images provided by an expert. As a case study, for each image modalities (ultrasound and x-ray), six test images are used for performance investigation. Details about the proposed algorithm and also conducted experiments are provided.