Fuzzy based residual shufflenet based breast cancer detection using mammogram images

Abstract Breast cancer is a condition in which abnormal breast cells grow and divide uncontrollably, resulting in the development of cancerous tissue. Detecting breast cancer at the preliminary phase is vital for improving survival rates and diagnostic outcomes. Though numerous Deep Learning (DL) frameworks have been proposed for detecting breast cancer, attaining accurate outcomes remains challenging. Thus, the Fuzzy-based Residual-ShuffleNet (Fuzzy RS-Net) is proposed for detecting breast cancer using mammogram images. Primarily, the mammogram image sourced from the dataset is fed as input to the wavelet domain filtering for noise reduction in the pre-processing stage. Further, the Object Segmentation Network (O-SegNet) is employed to segregate the affected region. Afterwards, image augmentation is effectuated using processes, such as rotation, shifting, and erasing. Further, feature extraction is processed, where texture features, Pyramid Histogram of Oriented Gradients (PHOG), and Weber Local Binary Pattern (WLBP) are extracted. At last, Fuzzy RS-Net is exploited for detecting breast cancer. In addition, the Fuzzy RS-Net is formulated by merging the Deep Residual Network (DRN), the Fuzzy concept, and ShuffleNet. Moreover, the developed Fuzzy RS-Net recorded the highest value of accuracy at 94.999%, sensitivity at 95.899%, and specificity at 93.889%.