Integration of Deep Learning Algorithms for Breast Cancer Detection using MRI Images with an STM32 Board

Breast cancer patient prognosis and treatment effectiveness are greatly enhanced by early identification and categorization. This research project focuses on analysing magnetic resonance imaging (MRI) images and conducting an extensive comparative investigation of five distinct deep learning algorithms to enhance the accuracy of breast cancer diagnosis. The goal is to evaluate the effectiveness and efficacy of deep learning models in correctly classifying instances of breast cancer. Different methods for identifying breast cancer using MRI, utilizing the CNN, LSTM, VGG16, and Autoencoder algorithms, are compared based on metrics such as accuracy, error rate, Classifier Success Index (CSI), F1score, Memory usage, and Execution Time. These metrics help understand their performance. In the comparative study involving five neural network architectures like CNN,LSTM,MobileNetV1,Auto-Encoder and VGG16, the Convolutional Neural Network(CNN) achieved the highest accuracy across the dataset.Due to constraints within the Edge Impulse platform,which supports certain architectures like MobileNet.So,MobileNetV1 was deployed for practical Implementation.Deployed MobileNetvV1,suitable for the STBL475EIOT01A2 Board,achieving an accuracy of 90.57% and 87.50% on the system(Intel Core i3).In conclusion, a comparison was made on execution time and memory usage between System and board.