A hybrid deep learning approach for breast cancer detection using AlexNet-ResNet with CNN features

Abstract

Breast cancer detection is a serious challenge in medical imaging, and use of deep learning techniques has shown promising results in improving accuracy and efficiency. This research introduces a hybrid deep learning strategy for breast cancer identification, utilizing the combination of AlexNet-ResNet architectures with convolutional neural network (CNN) features. The proposed model extracts intricate features from medical images and uses skip connections to handle deeper neural networks. The work introduces a hybrid deep learning approach for breast cancer detectionby proposing a hybrid model, called AlexNet-ResNet. Hybrid approach improves the network's generalization ability and provides a better understanding of key features contributing to breast cancer diagnosis. The AlexNet architecture possesses a remarkable depth of 8 layers, allowing it to excel in feature extraction when compared to the other models. On the other hand, ResNet architecture is a smart approach that not only considerably diminishes training time, but also elevates accuracy.The performance of the presented strategy was estimated by employing multiple metrics, such as specificity, accuracy, and sensitivity. Findings show that the AlexNet-ResNet approachgained impressive scores of 95.2% accuracy, 97.9% sensitivity, and 93.9% specificity. The proposed method could enhance the effectiveness of breast cancer screening programs and support personalized treatment planning.