Towards Comprehensive Breast Cancer Detection: Fusion of Dense Layer Model Features
Gadawalsa Kishor Kumar, Debendra Muduli, Kshirabdi Tanaya Rath, Durbar Sundar Ray, Santosh Kumar Sharma · 2024
For an early diagnosis and successful treatment, ultrasonography image analysis is essential for identifying breast cancer. In this discipline, deep learning has become a powerful tool that has demonstrated notable success in several medical imaging tasks,regarding the identification of cancer in the breast. In order to achieve accurate identification, our suggested model combines characteristics from three pre-trained Convolutional Neural Network (CNN) models: DenseNet201, DenseNet169, and DenseNet121. In this study, we use ImageNet weights and ultrasound image datasets to evaluate these models’ performance. After extensive testing, our model reached 94.57% accuracy, which is astounding. The model’s performance was evaluated through the use of confusion matrices and metrics such as precision & F1-score, which we determined using the Adam optimizer. The efficacy of deep learning-based techniques in identifying breast cancer from ultrasound pictures is highlighted by this study. It illustrates makes major contributions to medical analysis and helps discover breast cancers early.