Enhanced breast cancer detection and classification via CAMR-Gabor filters and LSTM: A deep Learning-Based method

Abstract

Breast cancer detection and classification are crucial for early diagnosis and effective treatment planning. This work proposed Modified Context-Aware Multiresolution Gabor Filters-Based Breast Cancer Classification (CAMR- GF-BCC) for identifying and categorizing breast cancer. Initially, the mammographic images are preprocessed through normalization and Gaussian filtering to enhance image quality and suppress noise. Subsequently, the processed images are segmented using the DeepLabv3 + model, which effectively delineates the regions of interest. Post-segmentation, the images are masked to isolate the significant features for analysis. Arithmetic features and CAMR-GF are then performed from these masked images, capturing essential characteristics pertinent to breast cancer detection. These features serve as inputs to a Long Short-Term Memory (LSTM). It is utilized in the work of categorization, leveraging its capability to handle sequential data and capture complex patterns. The proposed method is rigorously evaluated using standard performance metrics, showing its effectiveness in precisely identifying and categorizing breast cancer. This CAMR-GF-BCC work has 99.48 accuracy, 99.64 sensitivity, 99.14 specificity, 99.01 precision, and 98.16 F1-score. The results indicate a promising improvement in diagnostic accuracy, potentially aiding in timely and precise breast cancer treatment decisions.