ABSTRACT
This becomes a global issue, as breast cancer currently is one of the most urgent and dangerous diseases that require fast diagnostic procedures. The study revealed that advanced innovative machine learning algorithms could classify breast cancer from ultra-sound images. Image data set was containing 10000 samples, with benign, malignant and normal cells. Preprocessing resizing, normalization and noise reduction. We deployed four machine learning models that have been proven to be reliable: Inception V3, DenseNet 121, ResNet 50 and Convolutional Neural Network (CNN). CNN had the best overall performance with an accuracy of 95% and remarkable precision, recall, f1 score values. Other aimed deep learning models like DenseNet 121, ResNet 50 and Inception V3 can also be achieved good performance with more than 88% ranges of the accuracy. “Our models could set the stage for more successful cancer screening and improved patient outcomes, precisely characterising different types of breast cancer in ultrasound images” he said.
