Performance Metrics for Segmentation Algorithms in Brain MRI for Early Detection of Autism

Abstract

Autism is an abnormal condition of human brain neurons, which makes individuals attention deficient, unable to speak and several other neurodevelopmental disorders as detected in the children with the age group of 2 to 5 years. However, autism is a neurological irregularity with more than one behavioral problem. Autism would be generally detected by behavioral symptoms, but early detection was not possible with behavioral approach. So, studying the structure of brain by using MRI image of the brain would be an efficient technique in early detection of autism. Various image classification and segmentation methods have been developed by many researchers. This work proposes a new performance metrics to find out efficiency of segmentation algorithms.