Title: Research on the cancer cell's recognition algorithm based on the combination of competitive FHNN and FBPNN

Authors: Hu Qi; Duan Jin; Zhai Di

Addresses: Department of Electronics and Information Engineering, Changchun University of Science and Technology, Changchun, Jilin, China; Information Engineering Department Jilin Business and Technology College, Changchun, Jilin, China ' Department of Electronics and Information Engineering, Changchun University of Science and Technology, Changchun, Jilin, China ' Department of Public Security of Jilin Province, Changchun, Jilin, China

Abstract: In the field of traditional medicine, diagnosis on the microscopic image of cancer cells fallen into peritoneal effusion can often be recognised by the naked eye observation from personal experience, but uncertain factors such as intake angle and method of microscopic image will lead to frequent phenomenon such as misdiagnosis or cases that cannot be confirmed. A competitive fuzzy Hopfield neural network (FHNN) automatic clustering segmentation is used in this paper to segment suspected cells and nucleus from the microscopic image of peritoneal effusion exfoliated cells and to extract 15 features of cancer cells. Flexible BP neural network (FBPNN) is used in this paper as classifier that classifies and recognises the peritoneal effusion exfoliated cells. By analysing clinical cases, the proposed algorithm has been proved that it can acquire higher accuracy of cancer cells diagnosis.

Keywords: image segmentation; image recognition; competitive FHNN; fuzzy Hopfield neural networks; flexible BP neural networks; FBPNN; back propagation; automatic clustering; cancer cells; peritoneal effusion; feature extraction; classification accuracy; cancer diagnosis; medical images.

DOI: 10.1504/IJCSM.2016.077866

International Journal of Computing Science and Mathematics, 2016 Vol.7 No.3, pp.229 - 238

Received: 09 Apr 2016
Accepted: 25 Apr 2016

Published online: 17 Jul 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article