Research and Implementation of CAD on the Pathological Slices of Gastric Adenocarcinoma
|School||Xi'an University of Electronic Science and Technology|
|Course||Computer System Architecture|
|Keywords||Threshold segmentation Nucleus segmentation Watershed transform GVF Snake model Discriminant analysis|
Through the computer, medical image processing and analysis is an important application of image processing technology, combined with the characteristics of medical imaging . The computer-aided diagnosis system for medical research and clinical diagnosis , especially of great significance in the case of lack of experts . Existing computer-aided diagnosis of cancer are basically smears , for the most common pathological cell image , its organizational complexity , difficulty in cell division , has been little research . Gastric adenocarcinoma biopsy computer-aided diagnosis system . The composite cell image pre - processing technology, the composition of the nucleus contour segmentation algorithm and nuclear feature extraction and discrimination is the main work of this paper . Image preprocessing stage , according to the features of complex tissue biopsy of gastric cancer cell image analysis of pretreatment technology , composite pretreatment technology based on threshold segmentation to eliminate the impact of other organizations . In the nucleus segmentation stage , the combination of algorithms based on the the watershed transform the the ( watershed transformation ) GVFSnake model ( GVF Snake model ) . Watershed transform the initial contour of the nucleus , and then use GVF Snake model approximation of the true contours . In the nucleus feature extraction and determination phase , to give up the idea of a common , try to use statistically Fisher discriminant analysis of large sample data . Iterative screening and solving the characteristic parameter coefficient characteristic parameters , the eventual establishment of a cell sorting Fisher discriminant function expression . Experimental results show that this paper based on the improved gastric adenocarcinoma biopsy established computer-aided diagnosis system with the ability of cell division and better cell classification and discrimination ability to play a role in the pathological diagnosis for improving the reading level of junior doctors has a certain value . In this study, in theory enrich the contents of the computer-aided diagnosis has laid a good foundation for further study .