結腸鏡檢查是一種用來檢測大腸息肉的方法，可以幫助人們提早預 防癌變。然而，息肉形狀多樣，外型與腸道周邊黏膜組織相似，要準確 的檢測出息肉所在的位置及形狀，需消耗大量時間及人力，而且也常有 誤判的情況發生。
本篇論文提出一種新的深度學習模型架構。使用 Transformer 與 CNN 作為模型的雙骨架，訓練出從低階到高階不同層級的特徵，並使用 Fusion Block 以及 Cross-Features Attention 整合上述特徵來提高準確率。 其中 Fusion Block 負責處理低階層的特徵，Cross-Features Attention 則負 責將處理過的低階特徵及高階特徵做一個注意力機制的結合。所使用的 訓練集為 Kvasir-SEG、CVC-ClinicDB，測試集為 Kvasir-SEG、CVCClinicDB、CVC-ColonDB、ETIS、EndoScene。實驗結果顯示本篇論文 的方法能更好的預測出息肉的位置及形狀。

Colonoscopy is considered to be an effective technique for detecting colon polyps. It can help people to prevent cancer at an early stage. However, polyps have various shapes, the boundary between the polyp and the surrounding mucosa is not obvious, it usually takes a lot of time and labor to accurately detect the location and the shape of polyps, also, misclassification often occurs.
This Thesis proposes a deep learning model that uses Transformer and CNN as the backbones of the model, trying to train different levels of features from lower-level to higher-level, then integrate these features by using Fusion Block and Cross-Features Attention. Fusion Block is responsible for processing low-level features, while Cross-Features Attention is responsible for combining the processed lower-level features and higher-level features. The training sets are Kvasir-SEG, and CVC-ClinicDB, and the testing sets are Kvasir-SEG, CVCClinicDB, CVC-ColonDB, ETIS, and EndoScene. The experimental results show that the proposed method could perform better prediction.

[1] Health Promotion Administration Ministry of Health and Welfare, National Cancer Incidence Data, January 2022.

[2] Fatima A. Haggar, and Robin P. Boushey, Colorectal Cancer Epidemiology: Incidence, Mortality, Survival, and Risk Factor. Clin Colon Rectal Surg 2009, 191-197, 2009.

[3] Takahisa M., et al., Advances in image enhancement in colonoscopy for detection of adenomas. Nature Reviews Gastroenterology & Hepatology Volume 14, 305–314, 2017.

[4] Sang Bong A. and Dong Soo H. The Miss Rate for Colorectal Adenoma Determined by Quality-Adjusted, Back-to-Back Colonoscopies. Gut and Liver, Vol. 6, No. 1, 64-70, 2012.

[5] Katharina Z., F., and Susanne, S., Right-Sided Location Not Associated With Missed Colorectal Adenomas in an Individual-Level Reanalysis of Tandem Colonoscopy Studies. Gastroenterology Volume 157, Issue 3, September 2019, 660-671, 2019.

[6] Alexander V., and Mamonov Isabel N., Automated Polyp Detection in Colon Capsule Endoscopy. IEEE Transactions on Medical Imaging, 1488-1502, 2014.

[7] Marcelo Fiori P., and Guillermo S., A Complete System for Candidate Polyps
Detection in Virtual Colonoscopy. International Journal of Pattern Recognition and
Artificial Intelligence Vol. 28, No. 07, 1460014, 2014.

[8] Nima T., Automated Polyp Detection in Colonoscopy Videos Using Shape and
Context Information. IEEE Transactions on Medical Imaging, 630 - 644, 2016.

[9] Omid Haji M., Superpixel Based Segmentation and Classification of Polyps in
Wireless Capsule Endoscopy. 2017 IEEE Signal Processing in Medicine and Biology
Symposium (SPMB), 2017.

[10] Mojtaba A., and Majid M., Polyp Segmentation in Colonoscopy Images Using Fully Convolutional Network. 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2018.

[11] Yun Bo Guo, and Bogdan J. M., GIANA Polyp Segmentation with Fully
Convolutional Dilation Neural Networks. 14th International Joint Conference on
Computer Vision, Imaging and Computer Graphics Theory and Applications, 25 - 27, 2019.

[12] Brandao, P., Fully convolutional neural networks for polyp segmentation in
colonoscopy. Proceedings Volume 10134, Medical Imaging 2017: Computer-Aided
Diagnosis; 101340F (2017), 2017.

[13] Olaf R., Philipp F., and T., U-Net: Convolutional Networks for Biomedical Image
Segmentation. Medical Image Computing and Computer-Assisted Intervention –
MICCAI 2015, 234–241, 2015.

[14] Zongwei Z., UNet++: A Nested U-Net Architecture for Medical Image Segmentation. Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support, 3–11, 2018.

[15] Debesh J., ResUNet++: An Advanced Architecture for Medical Image Segmentation. IEEE International Symposium on Multimedia (ISM), 2019.

[16] Debesh J., and Pia H. S., A Comprehensive Study on Colorectal Polyp Segmentation With ResUNet++, Conditional Random Field and Test-Time Augmentation. IEEE Journal of Biomedical and Health Informatics ( Volume: 25, Issue: 6, June 2021), 2029 - 2040, 2021.

[17] Alex Sherstinsky., Fundamentals of Recurrent Neural Network (RNN) and Long
Short-Term Memory (LSTM) Network. Physica D: Nonlinear Phenomena Volume
404, March 2020, 132306, 2020.

[18] Deng-P. F., PraNet: Parallel Reverse Attention Network for Polyp Segmentation.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2020,
263–273, 2020.

[19] Ange L., CaraNet: Context Axial Reverse Attention Network for Segmentation of Small Medical Objects. Proceedings Volume 12032, Medical Imaging 2022: Image Processing; 120320D (2022), 2022.

[20] Nikhil K. T., DDANet: Dual Decoder Attention Network for Automatic Polyp
Segmentation. ICPR 2021: Pattern Recognition. ICPR International Workshops and
Challenges, 307–314, 2021.

[21] Lukas L., and Marco K., Auxiliary Tasks in Multi-task Learning. arXiv:1805.06334 , 2018.

[22] Sanghyun W., Jongchan P., CBAM: Convolutional Block Attention Module.
Proceedings of the European Conference on Computer Vision (ECCV), 2018, 3-19,
2018.

[23] Qibin H., Coordinate Attention for Efficient Mobile Network Design. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, 13713-13722, 2018.

[24] Jie H., Squeeze-and-Excitation Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018, 7132-7141, 2018.

[25] Ashish V., Noam S., Niki P., Jakob U., and Llion J., Attention Is All You Need.
Advances in Neural Information Processing Systems 30 (NIPS 2017), 2017.

[26] Alexey D., and Lucas B., An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. arXiv:2010.11929, 2020.

[27] Jieneng C., Yongyi L., TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation. arXiv:2102.04306, 2021.

[28] Yundong Z., Huiye L., TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation. Medical Image Computing and Computer Assisted
Intervention – MICCAI 2021, 14–24, 2021.

[29] Bo D., Polyp-PVT: Polyp Segmentation with Pyramid Vision Transformers.
arXiv:2108.06932, 2021.

[30] Jacob D., and Ming-W., BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. arXiv:1810.04805, 2018.

[31] Satyam M., FusAtNet: Dual Attention-based SpectroSpatial Multimodal Fusion
Network for Hyperspectral and LiDAR Classification.Proceedings of the IEEE/CVF
Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2020,
92-93, 2020.

[32] Wenhai W., Enze X., Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions.Proceedings of the IEEE/CVF International
Conference on Computer Vision (ICCV), 568-578, 2021.

[33] Wenhai W., Enze X., PVTv2: Improved Baselines with Pyramid Vision Transformer. Computational Visual Media volume 8, 415–424, 2022.

[34] Zhe W., Li S., and Qingming H., Cascaded Partial Decoder for Fast and Accurate
Salient Object Detection. Proceedings of the IEEE/CVF Conference on Computer
Vision and Pattern Recognition (CVPR), 3907-3916, 2019.

[35] Chen-Yu L., Saining X., Deeply-Supervised Nets. Proceedings of the Eighteenth
International Conference on Artificial Intelligence and Statistics, 562-570, 2015.

[36] Shih-En W., Convolutional Pose Machines. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 4724-4732, 2016.

[37] MA Y., LIU Q., QIAN Z., Automated Image Segmentation Using Improved PCNN
Model Based on Cross-entropy. Proceedings of 2004 International Symposium on
Intelligent Multimedia, Video and Speech Processing, 2004.

[38] Jiahui Y., UnitBox: An Advanced Object Detection Network. Proceedings of the 24th ACM international conference on Multimedia, 516–520, 2016.

[39] Debesh J., Kvasir-SEG: A Segmented Polyp Dataset. MMM 2020: MultiMedia
Modeling, 451–462, 2019

[40] N. Tajbakhsh, S. R. Gurudu, and J. Liang, “Automated polyp detection in
colonoscopy videos using shape and context information,” IEEE TMI, vol. 35, no. 2,
630–644, 2015.

[41] D. Vazquez, J. Bernal, F. J. S anchez, G. F., A. M. Lopez, A. Romero, M. Drozdzal,
and A. Courville, A benchmark for endoluminal scene segmentation of colonoscopy images, JHE, vol.2017, 2017.

[42] J. Silva, A. Histace, O. Romain, X. Dray, and B. Granado, Toward embedded
detection of polyps in wce images for early diagnosis of colorectal cancer, IJCARS,
vol. 9, no. 2, 283–293, 2014.

[43] Pogorelov, K., Randel, K., Griwodz, C., Sigrun, E., Lange, T., Johansen,D.,
Spampinato, C., Dang-Nguyen, D., Lux, M., Schmidt, P., Riegler, M.,Halvorsen, P.:
Kvasir: A Multi-Class Image Dataset for Computer AidedGastrointestinal Disease
Detection. In: Proceedings of Multimedia SystemsConference (MMSYS), 164–169.
ACM, 2017.

[44] Debesh J., Kvasir-SEG: Real-Time Polyp Detection, Localization and Segmentation in Colonoscopy Using Deep Learning. IEEE Access (Volume: 9), 40496-40510, 2021.

[45] Nikhil K. T., DDANet: Dual Decoder Attention Network for Automatic Polyp
Segmentation. ICPR 2021: Pattern Recognition. ICPR International Workshops and
Challenges, 307–314, 2021.

[46] Fan, D.P., Cheng, M.M., Liu, Y., Li, T., Borji, A.: Structure-measure: A new way to
evaluate foreground maps. In: IEEE ICCV, 4548–4557, 2017.

[47] Fan, D.P., Gong, C., Cao, Y., Ren, B., Cheng, M.M., Borji, A.: Enhanced-alignment
Measure for Binary Foreground Map Evaluation. In: IJCAI, 2018.

[48] T. Kim, H. Lee, and D. Kim, “Uacanet: Uncertainty augmented context attention for polyp segmentation,” in ACM MM, 2167-2175,2021.

[49] Fausto M., V-Net: Fully Convolutional Neural Networks for Volumetric Medical
Image Segmentation. 2016 Fourth International Conference on 3D Vision (3DV),
2016.

[50] Yanda M., Hongrun Z., Graph-Based Region and Boundary Aggregation for
Biomedical Image Segmentation. IEEE Transactions on Medical Imaging, 690-710,
2022