Breast Cancer Mammography Classification Using Convolutional Neural Networks and WOA-MPA Optimization
Saradha S A.Devendhiran · Journal of Propulsion Technology · 2023
Breast cancer analysis is crucial for female well-being and mortality statistics. Digital mammograms improve accuracy and survival rates. This study introduces a deep-learning algorithm-based training approach for automated breast cancer identification at early stages, enhancing edge detail and reducing false positives.The proposed methodology involves the integration of a Convolutional Neural Network (CNN) with an optimization strategy to establish a classification model for the diagnosis of breast cancer. The current study utilized a hybrid approach involving the Marine Predators Algorithm (MPA) and the Whale Optimization Algorithm (WOA) to determine the well-suitable hyperparameter values for the CNN framework. The proposed approach employs Inception v3, a previously trained convolutional neural network model, and DenseNet. MPA-WOA with Inception v3 and MPA-WOA with DenseNet are two instances of architectu res that integrate this paradigm with the MPA-WOA algorithm. By conducting a comparative evaluation of the attainment of two hybrid models, this study has established that the MPA-WOA with DenseNet has demonstrated an exactness rate of 94% and 95% for the CBIS-DDSM and MIAS datasets, respectively.