Mammogram Image Segmentation Using Hybridization of Fuzzy Clustering and Optimization Algorithms

Abstract

Mammogram images have the ability to assist physicians in detecting breast cancer caused by cells abnormal growth. But due to visual interpretation, false results can be obtained. In this paper, to reduce false results, image segmentation is carried out to find breast cancer mass. Image segmentation using Fuzzy clustering: K means, FCM, and FPCM shows result better than other existing methods but initialization problem and sensitivity to noise do not make them to achieve better accuracy. Various extension of the FCM for segmentation is developed. But most of them modify the objective function which changes the basic FCM algorithm. Hence efforts have been made to develop FCM algorithm without modifying objective function for better segmentation. We have proposed a technique GA-ACO-FCM, which is the hybridization of optimization tools: genetic algorithm and ant colony optimization with fuzzy C means .GA-ACO-FCM is suitable to overcome initialization problem of FCM and shows better results with achieving high accuracy.

Towards Better Segmentation of Abnormal Part in Multimodal Images Using Kernel Possibilistic C Means Particle Swarm Optimization With Morphological Reconstruction Filters

“…Combination of Genetic Algorithm -Ant Colony optimization-Fuzzy C means (GA-ACO-FCM) algorithms (Guru Kalyan Kanungo et al, 2014) aids in segmenting the breast tumor part and achieves high accuracy when compared with other existing techniques like GA and ACO. Weighted PSO optimization (Seetha & Santhosh baboo, 2017) applied to extract the tumor part of mammogram breast cancer images and classifies the micro calcifications into benign, malignant or normal images.…”

Research and Review on Novel Spread Spectrum Communication Theory

“…Spread spectrum and frequency hopping technology of antijamming mechanism is different, have their own characteristics and shortcomings. In terms of strong fixed frequency interference, anti-interference is spread through the correlation algorithm to obtain the processing gain to achieve anti-jamming, but more than the fixed frequency interference tolerance will lead to a precipitous decline in spread spectrum communication interrupt or performance of the system [7]; And frequency hopping system is adopted to avoid the method of anti-interference, strong fixed frequency interference can only be one or several frequency interference frequency hopping system, will affect the performance of the system is not very serious. The detailed discussion on the theoretical parts of the method is conducted in the next section.…”

Mammography Images Segmentation via Fuzzy C-mean and K-mean

“…A. Elmoufidi, K. El Fahssi, et al [13] shows a strategy for segment and distinguishes the limit of various breast tissue districts in mammograms by utilizing Seed Based Region Growing (SBRG) method and dynamic K-means clustering algorithm. G. K. Kanungo, et al [14] to diminish false outcomes, image segmented is carried out for discover breast cancer mass. Image segmentation utilizing Fuzzy clustering: FCM, K means, and FPCM demonstrates result superior to other existing methods but sensitivity to noise don't improve them to accomplish exactness.…”