Method for automatically segmenting brain micro-hemorrhagic spots

• Main Author: BIAN YUEYAN
• Format: Patent
• Language: Chinese; English
• Published: 03.09.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention provides a method for automatically segmenting brain micro-hemorrhagic spots, which is characterized by comprising the following five steps of: 1, preprocessing data, including reading an SWI image and a QSM image; extracting brain parenchyma, and removing tissues except the brain parenchyma; 2, extracting a micro-hemorrhage candidate frame, including extracting candidate micro-hemorrhage points through a self-adaptive threshold value; performing morphological processing on the candidate micro bleeding points; screening candidate micro bleeding points according to the volume; 3, classifying and identifying candidate micro-hemorrhage, including normalizing micro-hemorrhage points; training a classification network; inputting data in the micro-hemorrhage candidate frame into the trained classification network for classification; and counting the volume and the number of the micro-hemorrhage in the micro-hemorrhage candidate box. According to the method, the SWI image and the QSM image are utilized