Enhanced Breast Cancer Classification: A Novel Fusion of Deep Features, Shape Features and GCLM Features

Early and accurate detection of breast cancer is essential due to its significant impact on women’s health. This study presents a hybrid classification approach that combines deep learning features with traditional handcrafted features to improve the performance of breast cancer diagnosis, particularly for small datasets. Deep learning features were extracted using the ResNet50 model, while handcrafted features were derived from shape descriptors and Gray-Level Co-Occurrence Matrix (GLCM) characteristics. These features were fused and evaluated using three machine learning classifiers: Support Vector Machine (SVM), Decision Tree, and Random Forest. The SVM classifier achieved the highest classification accuracy of 96.88%, outperforming other models. Additionally, the SVM model demonstrated superior precision (96.54%), recall (97.12%), and F1-score (96.83%), highlighting its ability to accurately differentiate between benign and malignant cases. While LeNet was also evaluated, the combination of ResNet50 with handcrafted features consistently yielded better results, underscoring the effectiveness of this hybrid approach in breast cancer classification.