ColoEnsemCADx: A Deep Ensemble Framework for Colorectal Disease Diagnosis from Colonoscopy Images
Keywords:
Colonoscopy images, Colorectal cancer, Deep learning, Ensemble model, Explainable AIAbstract
Colorectal cancer ranks among the most common types of cancer. To enhance the prognosis of patients suffering from it, an effort must be made to improve the precision and the timeline of the diagnoses. Colonoscopy, in particular, is a competent technique, although it is subjective and resource-heavy. We, therefore, designed ColoEnsemCADx, which is an interpretable, multi-tiered deep learning framework that automates the classification of colorectal cancer with colonoscopic images obtained from the Kvasir dataset. The set comprises 4,000 images spread across eight classes, which was further augmented to 5,600 images for balance. At stage one, three hybrid CNN architectures-RDV-2025, IEM-2025, and DRE-2025 were engineered with pre-trained models to different feature regions to ensure diversity in the extracted features. At stage two, decision boundaries were further enhanced by feature refinement through XGBoost. Multi-class SVM was used for the final classification in stage three. Testing accuracy, F1 score, and AUC increased and surpassed existing benchmarks, achieving results of 99.21%, 98.43%, and 98.16% respectively with the DRE-2025 + XGBoost + SVM ensemble. Due to its interpretability and high performance, ColoEnsemCADx can be a trustworthy diagnostic assistant in real-world healthcare settings.
References
M. K. Fadafen, and K. Rezaee, “Ensemble-based multi-tissue classification approach of colorectal cancer histology images using a novel hybrid deep learning framework,” Scientific Reports, vol. 13, no. 8823, May 2023, doi: https://doi.org/10.1038/s41598-023-35431-x https://doi.org/10.1038/s41598-023-35431-x
W. S. Liew, T. B. Tang, C.-H. Lin, and C.-K. Lu, “Automatic colonic polyp detection using the integration of modified deep residual convolutional neural network and ensemble learning approaches,” Computer Methods and Programs in Biomedicine, vol. 206, p. 106114, Jul. 2021, doi: https://doi.org/10.1016/j.cmpb.2021.106114
F. Younas, M. Usman, and W. Q. Yan, “An ensemble framework of deep neural networks for colorectal polyp classification,” Multimedia Tools and Applications, vol. 82, pp. 18925–18946, Nov. 2022, doi: https://doi.org/10.1007/s11042-022-14177-0
A. Karthikeyan, S. Jothilakshmi, and S. Suthir, “Colorectal cancer classification based on deep ensemble model with self-adaptive training model,” International Journal of Biomedical Engineering and Technology, vol. 44, no. 4, pp. 367–396, Apr. 2024, doi: https://doi.org/10.1504/ijbet.2024.138065
Q. Ke, W.-S. Yap, Y. K. Tee, Y. C. Hum, H. Zheng, and Y.-J. Gan, “Advanced deep learning for multi-class colorectal cancer histopathology: integrating transfer learning and ensemble methods,” Quantitative Imaging in Medicine and Surgery, vol. 15, no. 3, pp. 2329–2346, Mar. 2025, doi: https://doi.org/10.21037/qims-24-1641
S. Ghosh, A. Bandyopadhyay, S. Sahay, R. Ghosh, I. Kundu, and K. C. Santosh, “Colorectal histology tumor detection using ensemble deep neural network,” Engineering Applications of Artificial Intelligence, vol. 100, p. 104202, Apr. 2021, doi: https://doi.org/10.1016/j.engappai.2021.104202
M. S. Hossain et al., “DeepPoly: Deep learning-based polyps segmentation and classification for autonomous colonoscopy examination,” in IEEE Access, vol. 11, pp. 95889-95902, 2023, doi: https://doi.org/10.1109/ACCESS.2023.3310541
W. Wang, Q. Li, C. Xiao, D. Zhang, L. Miao, and L. Wang, “An improved boundary-aware U-Net for ore image semantic segmentation,” Sensors, vol. 21, no. 8, p. 2615, Apr. 2021, doi: https://doi.org/10.3390/s21082615
K. E. Delowar, M. B. Uddin, Md. Khaliluzzaman, R. I. Rabbi, Md. J. Hossen, and M. M. Hossen, “PolyNet: A self-attention based CNN model for classifying the colon polyp from colonoscopy image,” Informatics in Medicine Unlocked, vol. 56, pp. 101654–101654, May 2025, doi: https://doi.org/10.1016/j.imu.2025.101654
M. Liu, Y. Pei, M. Wu, and J. Wang, “Focal cosine-enhanced efficientNetB0: A novel approach to classifying breast histopathological images,” Information, vol. 16, no. 6, pp. 444–444, May 2025, doi: https://doi.org/10.3390/info16060444
A. Kajala, S. Jaiswal, and R. Kumar “Breaking the black box: Heatmap-driven transparency to breast cancer detection with Efficientnet and Grad CAM,” Educational Administration: Theory and Practice, vol. 30, no. 5, pp. 4999-5009, Available: https://kuey.net/index.php/kuey/article/view/3738/2461
S. M. Fati, E. M. Senan, and A. T. Azar, “Hybrid and Deep Learning Approach for Early Diagnosis of Lower Gastrointestinal Diseases,” Sensors, vol. 22, no. 11, p. 4079, May 2022, doi: https://doi.org/10.3390/s22114079
A. S. N. Raju, K. Jayavel, T. Rajalakshmi, and M. Rajababu, “CRCFusionAICADx: Integrative CNN-LSTM Approach for accurate colorectal cancer diagnosis in colonoscopy images,” Cognitive Computation, vol. 17, no. 14, Nov. 2024, doi: https://doi.org/10.1007/s12559-024-10357-2
N. K. Tomar, D. Jha, K. Biswas and U. Bagci, “Transformer-enhanced iterative feedback mechanism for polyp segmentation,” ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India, 2025, pp. 1-5, doi: https://doi.org/10.1109/ICASSP49660.2025.10890567
C. C. Ukwuoma et al., “Enhancing histopathological medical image classification for Early cancer diagnosis using deep learning and explainable AI – LIME & SHAP,” Biomedical Signal Processing and Control, vol. 100, part C, pp. 107014–107014, Oct. 2024, doi: https://doi.org/10.1016/j.bspc.2024.107014
Akella et al., “Exploring vision transformers and XGBoost as deep learning ensembles for transforming carcinoma recognition,” Scientific Reports, vol. 14, no. 30052, Dec. 2024, doi: https://doi.org/10.1038/s41598-024-81456-1
