Performance and generalization analysis of machine learning, deep learning, and transformer models for histopathology image classification

Histopathology image classification plays a critical role in computer-aided diagnosis by supporting pathologists in disease detection and grading. With the rapid advancement of artificial intelligence, a wide range of machine learning, deep learning, and transformer-based models have been applied to histopathological image analysis. However, a systematic and fair comparison of these approaches under a unified experimental setting remains limited. In this study, we present a comprehensive performance and generalization analysis of classical machine learning classifiers, convolutional neural network (CNN) models, and vision transformer-based architectures for histopathology image classification. Publicly available benchmark datasets were used to evaluate the models using standardized preprocessing, training protocols, and evaluation metrics. The comparative analysis highlights the strengths and limitations of each category of methods in terms of classification accuracy, robustness, and computational complexity. Experimental results demonstrate that deep learning and transformer-based models consistently outperform traditional machine learning approaches, while transformer models show improved generalization capability on complex tissue patterns. The findings of this study provide practical insights for selecting suitable classification models in histopathology-based diagnostic applications and contribute to the development of reliable medical imaging decision-support systems.