Title: Mitosis detection from histological images using handcrafted features and artificial neural network
Authors: Hanan Hussain; Omar Hujran; K.P. Nitha
Addresses: Department of Information Technology, Ajman University, University Street, 2758, Ajman, United Arab Emirates ' Innovation in Government and Society Department, College of Business and Economics, United Arab Emirates University, Al-Ain, United Arab Emirates ' Department of Computer Science and Engineering, Vidya Academy of Science and Technology, Thrissur, India
Abstract: Mitosis is defined as the rapid division of cells and its count is relevant to predict the grading of breast cancer. Since manual mitosis detection is time consuming and prone to errors, a fast and accurate detection approach is proposed using handcrafted features with artificial neural network (ANN). This method includes three steps: 1) image pre-processing involves conversion on RGB image to red-channel; 2) segmentation which is done using fuzzy C-means clustering and handcrafted features are extracted; 3) classification in which both random forest classifier and ANN are ensemble to predict the outcome. The system was tested with Mitos-Atypia14 dataset and an accuracy of 91.6% is obtained.
Keywords: breast cancer detection; mitosis detection; artificial neural network; ANN; random forest classifier.
International Journal of Computer Aided Engineering and Technology, 2022 Vol.16 No.2, pp.240 - 256
Received: 21 Feb 2019
Accepted: 29 Apr 2019
Published online: 11 Feb 2022 *