🤖 AI Summary
This study addresses the lack of systematic comparison among convolutional neural networks (CNNs), vision Transformers (ViTs), and hybrid architectures for colorectal histopathological image classification. Leveraging the Kather dataset, the authors evaluate twelve prominent models—including ResNet34, ConvNeXt-Tiny, ViT-B/16, and EVA-02—under a unified pipeline of ImageNet pretraining, transfer learning, and fine-tuning. For the first time, the three major deep learning paradigms are comprehensively benchmarked under standardized conditions, with performance analyzed through accuracy, interpretability, and per-class metrics to identify strengths and challenging categories. Results show that all models achieve high accuracy ranging from 93.2% to 97.1%, with the hybrid model EVA-02 attaining the best performance at 97.1%. Vision Transformers generally outperform CNNs, while modern CNNs offer a favorable trade-off between accuracy and computational complexity.
📝 Abstract
Deep learning has become an important tool in computational pathology, enabling automated analysis of histopathological images. While convolutional neural networks (CNNs) have traditionally dominated this field, transformer-based and hybrid architectures have recently demonstrated promising performance. However, comprehensive comparisons of these approaches for colorectal histopathology remain limited. This study evaluated twelve ImageNet-pretrained CNN, transformer, and hybrid architectures using the Kather colorectal histopathology dataset containing 5,000 image tiles from eight tissue classes. All models were trained using a standardized transfer-learning and fine-tuning protocol and assessed using multiple performance metrics, including accuracy, precision, sensitivity, specificity, F1-score, ROC-AUC, Cohen's kappa, and Matthews correlation coefficient. All evaluated models achieved high classification performance, with accuracies ranging from 93.2% to 97.1%. EVA-02 achieved the highest overall performance (97.1% accuracy, 97.0% F1-score), closely followed by ViT-B/16. Among CNNs, ResNet34 and ConvNeXt-Tiny demonstrated highly competitive performance, achieving accuracies of 96.4% and 96.3%, respectively. Transformer architectures generally produced the strongest results across evaluation metrics, although the performance gap between the best transformer and CNN models was relatively small. Per-class analysis showed consistently strong classification performance across all tissue categories, with Complex Stroma representing the most challenging class. Overall, transformer-based architectures achieved the highest predictive performance, whereas modern CNNs provided a favorable balance between accuracy and model complexity. These findings provide a comprehensive benchmark of major deep learning paradigms for colorectal histopathology classification.