A Hybrid CNN-SVM Prediction Approach for Breast Cancer Ultrasound Imaging

This paper discusses the development of a hybrid Convolutional Neural Network (CNN)-Support Vector Machine (SVM) model for automated breast tumor detection using ultra-sound images. The study aims to compare the accuracy of the proposed model with the AlexNet CNN and develop a generalized image detection model that can detect tumors of all types across the human anatomy. The use of ultrasound imaging is justified based on its non-invasive nature and cost-effectiveness, and the possibility of collecting large datasets. The study reviews the prior work on deep learning and CNNs for tumor detection and segmentation in medical imaging, highlighting their potential for improving accuracy and efficiency. The findings of this study have the potential to improve the prognosis and treatment of cancer, a serious health condition affecting a significant number of people worldwide. CNN-SVM model had an accuracy of 91% and AlexNet at 88% with validation accuracy at 60% and 65% respectively. Showing the hybrid model is better in terms of accuracy and has more potential as a future base for medical image modeling prediciton.