Hybrid YOLO Feature Extraction with RF–AdaBoost Classifier for Accurate Breast Tumor Detection

The malignancy is the highest diagnosed breast cancer in India with one out of every 25.8 being breast cancer. The problem of tumor stages is a critical factor in treatment planning and improving patients survival rates due to accurate and early diagnoses. In this paper, there is a hybrid Cancer Classification Model (CCM) that uses ultrasound images which have been processed to detect and classify early on cancerous tumours as benign or malignant with estimation of the stages of disease progression and probability of survival.The proposed CCM uses YOLO to rapidly and precisely localize the tumors and uses a hybrid Random Forest (RF)-AdaBoost ensemble classifier to classify the tumors using robust features, and the result is to have high diagnostic accuracy at low computational complexity. Combining the feature extraction with deep learning and the ensemble learning enhances the quality of classification and supports clinical decision-making.The experiment conducted between the YOLO + RF–AdaBoost system and the traditional classifiers like Naive-Bayes and Decision Tree classifiers. The effectiveness of the proposed approach is proved by experimental results showing a high level of accuracy of 96 per cent, which proves the authoritative usage in detecting breast cancer at the initial early stages (I and II).