AN INTELLECTUAL DEEP LEARNING STRATEGY FOR BREAST CANCER DIAGNOSIS USING 3D TRANS-RESUNET3+ ASSISTED SEGMENTATION MECHANISM

Globally, breast cancer is a deadly disease for women. Early detection of breast cancer helps control and avoids high risk. The mammogram imaging approach is used to screen and detect abnormalities and assess their precise location. Better-quality mammogram images are highly required to detect and enhance the patient’s prognosis of breast cancer in an earlier stage. The traditional technique for breast screening considers X-ray images for the early detection of tumors. In dense breasts, the sensitivity analysis for breast cancer detection is generally a low value of 0.8261. Furthermore, it suffers from problems when screening the images to detect small abnormalities. Enhanced deep learning is an improved version of a machine learning mechanism that demands a higher computation power. Thus, this research focuses on designing an adaptive deep-learning model to identify tumor locations for better diagnosis. The required mammogram images are collected through standard data sources and fed to the segmentation mechanism to detect the exact location of cancerous tissue. Here, the Three-Dimensional Trans-ResUnet3[Formula: see text](3DTRUnet3[Formula: see text]) model is utilized for the effective abnormal region segmentation. The segmented images are fed into the classification network named Attention-based Adaptive ShufflenetV2 with Long Short-Term Memory (AASv2-LSTM) for diagnosing breast cancer. Hence, the tumor diagnosis is effectively done using the Best and Worst Fitness-aided Golf Optimization (BWF-GO) to improve classification performance. The diagnosis performance of the implemented task is validated with classical models concerning various measures. Throughout the entire experimental validation, the developed model obtains 96.92% accuracy which ensures the system’s effectiveness. This evaluation helps to provide a better diagnosis process that certainly improves the chance of survival and reduces the risk of mortality rate.