The paper "Adaptive Segmentation of MRI Data" by W. M. Wells III, W. E. L. Grimson, R. Kikinis, and F. A. Jolesz introduces a new method called *adaptive segmentation* to address the challenges of intensity-based classification in Magnetic Resonance Imaging (MRI). The method corrects and segments MRI images by using knowledge of tissue intensity properties and intensity inhomogeneities, leveraging the Expectation-Maximization (EM) algorithm for more accurate tissue type segmentation and better visualization of MRI data. The authors demonstrate the effectiveness of the method through a study involving over 1000 brain scans, showing comparable accuracy to manual segmentation and superior performance to supervised multi-variate classification in segmenting gray and white matter. The method is implemented for axial, coronal, and sagittal brain slices using conventional and surface coils, and its performance is validated through various tests, including comparisons with manual and supervised segmentations. The paper also discusses the method's advantages over existing techniques and its potential applications in medical imaging.The paper "Adaptive Segmentation of MRI Data" by W. M. Wells III, W. E. L. Grimson, R. Kikinis, and F. A. Jolesz introduces a new method called *adaptive segmentation* to address the challenges of intensity-based classification in Magnetic Resonance Imaging (MRI). The method corrects and segments MRI images by using knowledge of tissue intensity properties and intensity inhomogeneities, leveraging the Expectation-Maximization (EM) algorithm for more accurate tissue type segmentation and better visualization of MRI data. The authors demonstrate the effectiveness of the method through a study involving over 1000 brain scans, showing comparable accuracy to manual segmentation and superior performance to supervised multi-variate classification in segmenting gray and white matter. The method is implemented for axial, coronal, and sagittal brain slices using conventional and surface coils, and its performance is validated through various tests, including comparisons with manual and supervised segmentations. The paper also discusses the method's advantages over existing techniques and its potential applications in medical imaging.