Eel and grouper lyrebird optimizer based fractal deep spiking residual network for breast cancer detection using mammogram images

Breast cancer is the most common cancer type amongst women throughout the world. Mammography is a successful imaging modality for the earliest diagnosis of breast cancer in its beginning stages. Owing to less visibility and bad contrast in mammogram (MG) images, an earlier identification of breast cancer is a critical step for efficacious treatment of this disease. Here, the Eel and Grouper Lyrebird Optimizer-based Fractal Deep Spiking Residual Network (EGLO_FDSRN) is introduced for the detection of breast cancer using MG images. Initially, the MG image from mammographic image datasets is considered as input. The Wiener filter is utilized to pre-process the considered MG image. After that, the cancer region is segmented utilizing U-Next. Then, the cancer region segmented image is augmented by applying image augmentation methods, and thereafter, features suitable for the detection process are extracted. Lastly, breast cancer is detected by Fractal Deep Spiking Residual Network (FDSRN). However, FDSRN is developed by combining FractalNet with Spiking ResNet (S-ResNet). Furthermore, FDSRN is trained by Eel and Grouper Lyrebird Optimizer (EGLO), which is introduced by incorporating Eel and Grouper Optimizer (EGO) and Lyrebird Optimization Algorithm (LOA). In addition, EGLO_FDSRN acquired the best outcomes with 91.780% of accuracy, 90.689% of sensitivity, and 91.610% of specificity.