The early detection and accurate diagnosis of colorectal polyps play a crucial role in preventing colorectal cancer. Traditional computer-aided diagnosis methods faced the problem of insufficient accuracy when handling complex backgrounds, irregular shapes, and small polyps. This paper proposed an intelligent segmentation method for colorectal polyps based on a dual-channel adaptive network. The method designed a dual-channel feature enhancement module. It processed spatial and semantic features in parallel and enhanced the model's perception ability of polyp morphology. The method also constructed an adaptive feature fusion mechanism. It achieved dynamic integration of multi-scale features and improved the recognition accuracy of the polyp boundary. Additionally, a multi-level constrained learning strategy was proposed. It introduced boundary constraints and reliability assessment and enhanced the accuracy and reliability of the segmentation results. Experiments were carried out on a dataset with 612 colorectal endoscopic images. The Dice coefficient of the proposed method reached 0.912, which was 4.0 percentage points higher than that of existing methods. The detection rate of small polyps (diameter < 5mm) increased by 5.0 percentage points. This verifies the effectiveness of the method. This research provides a new technical solution for the intelligent diagnosis of colorectal polyps and has important clinical application value.

Yichen Guo, Lifeng He. Dual-Channel Adaptive Deep Network with Uncertainty Estimation for Colorectal Polyp Segmentation. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 7: 17-26. https://doi.org/10.25236/AJCIS.2025.080703.

[1] Ronneberger O, Fischer P, Brox T. U Net: Convolutional Networks for Biomedical Image Segmentation [C]// Proc of MICCAI. Munich: Springer, 2015: 234-241.

[2] Fan D P, Zhou T, Ji G P, et al. PraNet: Parallel Reverse Attention Network for Polyp Segmentation [C]// Proc of MICCAI. Shenzhen: Springer, 2020: 263-273.

[3] Chen J, Lu Y, Yu Q, et al. TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation [J/OL]. arXiv, 2021.(2021 02 11)[2021 02-11]. http://arxiv.org/abs/2102.04306.

[4] Jha D, Smedsrud P H, Riegler M A, et al. ResUNet++: An Advanced Architecture for Medical Image Segmentation [J]. IEEE Journal of Biomedical and Health Informatics, 2020, 24(11): 3278-3286.

[5] Oktay O, Schlemper J, Le Folgoc L, et al. Attention U Net: Learning Where to Look for the Pancreas [C]// Medical Imaging with Deep Learning (MIDL). London: PMLR, 2018.

[6] Dosovitskiy A, Beyer L, Kolesnikov A, et al. An Image is Worth 16×16 Words: Transformers for Image Recognition at Scale [C]// Proc of ICLR. New Orleans: OpenReview.net, 2021.

[7] Liu Z, Lin Y, Cao Y, et al. Swin Transformer: Hierarchical Vision Transformer using Shifted Windows [C]// Proc of ICCV. Montreal: IEEE, 2021: 10012-10022.

[8] He K., Zhang X., Ren S., Sun J. Deep Residual Learning for Image Recognition. CVPR, 2016.

[9] Tao Xiting, Ye Qing. Parallel Dual-Branch Skin Lesion Image Segmentation Integrating CNN and Transformer [J]. Computer Applications and Research, 2024, 41(8): 2554–2560.

[10] Gal Y, Ghahramani Z. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning [C]// Proc of ICML. New York: JMLR, 2016: 1050-1059.

[11] Kendall A, Gal Y. What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? [C]// Advances in Neural Information Processing Systems (NeurIPS). Long Beach: NeurIPS Press, 2017: 5574-5584.

[12] Lakshminarayanan B, Pritzel A, Blundell C. Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles [C]// Advances in Neural Information Processing Systems (NeurIPS). Long Beach: NeurIPS Press, 2017: 6402-6413.

[13] Kervadec H., Dolz J., Desrosiers C., Ayed I. B. Boundary Loss for Highly Unbalanced Segmentation. MICCAI, 2019.

[14] Park J, Woo S, Lee J. Adaptive Region Growing based Polyp Segmentation [J]. IEEE Transactions on Biomedical Engineering, 2017, 64(9): 2050- 2060.

[15] Liu H, Hu P, Zhang L. Graph Cut Based Colon Polyp Segmentation [J]. Journal of Biomedical Informatics, 2019, 95: 103-112.

[16] Silva J S, Pérez R, Bernal J, Sánchez F. Toward Benchmarking Polyp Detection Methods: CVC ClinicDB [C]// Proc of CVPR Workshops. Honolulu: IEEE, 2014: 1-5.

[17] Liang Liming, He Anjun, Dong Xin, et al. Colorectal Polyp Segmentation Algorithm Based on PVTv2 and Multi-Scale Boundary Aggregation [J]. Computer Applications and Research, 2023, 40(5): 1553–1558.

[18] Pogorelov K, Sanchez F, et al. Kvasir: A Multi Class Polyp Segmentation Dataset [C]// Proc of BMVC. London: BMVA Press, 2017.

[19] Shorten C, Khoshgoftaar T M. A Survey on Image Data Augmentation for Deep Learning [J]. Journal of Big Data, 2019, 6(60): 1-48.

[20] Kingma D P, Ba J. Adam: A Method for Stochastic Optimization [C]// Proc of ICLR. San Juan: OpenReview.net, 2015.

[21] Loshchilov I, Hutter F. SGDR: Stochastic Gradient Descent with Warm Restarts [C]// Proc of ICLR. Toulon: OpenReview.net, 2016.

[22] Zou K H, Warfield S K, Bharatha A, et al. Statistical Validation of Image Segmentation Quality Based on a Spatial Overlap Index [J]. Academic Radiology, 2004, 11(2): 178-189.

[23] Jaccard P. The Distribution of the Flora in the Alpine Zone [J]. New Phytologist, 1912, 11(2): 37-50.

[24] Powers D M W. Evaluation: From Precision, Recall and F Score to ROC, Informedness, Markedness & Correlation [J]. Journal of Machine Learning Technologies, 2011, 2(1): 37-63.

[25] Silva U T, et al. ETIS LARIB Polyp DB: A Large Scale Polyp Dataset [J]. Endoscopy, 2015, 47(7): 597-600.