A novel deep neural network with adaptive sine cosine crow search (DNN-ASCCS) model for content based medical image reterival

Article type: Research Article

Authors: Subha Darathy, C.^{a; *} | Agees Kumar, C.^b

Affiliations: [a] Department of CSE, Arunachala College of Engineering for Women, Vellichanthai | [b] Department of EEE, Arunachala College of Engineering for Women, Vellichanthai

Correspondence: [*] Corresponding author. C. Subha Darathy, Assistant Professor, Department of CSE, Arunachala College of Engineering for Women, Vellichanthai. E-mail: [email protected].

Abstract: Tumor is the second major cause of death in women worldwide. Breast cancer diagnosis and treatment can be difficult for radiologists. As a result, primary care helps to avoid disease and mortality. The study’s main goal is to improve treatment choices and to save lives by detecting breast cancer earlier. For classification problems, we propose a DNN-ASCC architecture in this study. The Fast Non-Local Means Filter completes the initial preprocessing stage. The binary grasshopper optimization algorithm (BGOA) and the grey-level run length matrix are utilized to choose the best features for the feature extraction operation. The suggested hybrid classifier (DNN-ASCCS) is critical for identifying normal and malignant tumors. Breast cancer is accurately detected by the suggested hybrid classifier. The recommended (DNN-ASCCS) was developed using MATLAB and datasets from the BIDCIDRI. The results of the simulation showed that the proposed technique has an accurate results in classification (99.17 percent) and robustness analysis is also done. When compared to alternative approaches, experimental results show that the suggested method is efficient.

Keywords: Breast cancer, DNN-ASCCS, content based medical image retrieval

DOI: 10.3233/JIFS-222872

Journal: Journal of Intelligent & Fuzzy Systems, vol. 44, no. 2, pp. 3427-3440, 2023