無線膠囊內視鏡是一種非侵入式方法，用於檢查整個小腸腸道。建立一個自動小腸瘜肉分割系統可以協助醫師診斷瘜肉，提高臨床表現，準確診斷病灶位置並節省專家診斷的時間。然而，由於無線膠囊內視鏡所拍攝到的瘜肉影像並不多，訓練深度學習模型變得困難。
資料增強是一種普遍用來解決資料量不足的方法，傳統的資料增強技術，如旋轉、翻轉、隨機裁剪，常用於解決資料不足的問題。這些方法雖然可以增加資料的數量，但沒辦法很有效地增加資料集的多樣性跟泛化性。因此，本篇論文提出了一個流程用於產生合成資料以增加資料集的資料量跟多樣性。
語意分割方法可以用來描繪出瘜肉的輪廓幫助精準診斷跟干預治療。我們使用合成數據去訓練一個針對無線膠囊內視鏡影像中小腸瘜肉分割任務的深度學習模型。此外，我們利用來自大腸鏡的大腸息肉資料集做遷移學習以加快模型的收斂速度。在合成數據跟遷移學習得幫助下，無線膠囊內視鏡影像中小腸瘜肉分割的模型達到了0.89的Dice係數，0.9的精確度和0.88的召回率。該模型能夠協助醫生在無線膠囊內視鏡診斷中進行小腸瘜肉的偵測。
在某些任務中，特別是醫療領域中的特定任務中，要收集足夠的具有標註資料用以訓練深度學習模型是具有挑戰性的。我們提出的流程為極少量資料的任務提供了一種資料增強的方法，有效地增加它們的訓練資料集。因此這個流程未來有潛力應用在其他小數據集任務中開發分割模型。

Wireless Capsule Endoscopy is a noninvasive method for examining the entire small intestine. An automatic polyp segmentation system can assist physicians in diagnosing polyps, improving clinical performance, accurately accessing lesions, and reducing expert time. However, it is challenging to collect enough polyp images captured by wireless capsule endoscopy. Due to the small dataset, training a deep learning model becomes difficult.
Traditional data augmentation techniques, such as rotating, flipping, and random cropping, are commonly used to address the problem of insufficient data. However, these techniques may not introduce enough diversity and generality to the training dataset. In this study, we propose a procedure for generating synthetic data to increase the amount of training data and enhance dataset diversity. Semantic segmentation methods can delineate the polyp regions to help accurately diagnose and intervene. We utilized the synthetic data to train a deep learning model for small bowel polyp segmentation in wireless capsule endoscopy images. Moreover, we leveraged transfer learning from a colon polyp dataset to accelerate the convergence of the segmentation model. We effectively utilized the synthetic data and employed transfer learning techniques to improve the performance of the small bowel polyp segmentation model in wireless capsule endoscopy image. The small bowel polyp segmentation model in wireless capsule endoscopy images achieved a Dice coefficient of 0.89, precision of 0.9, and recall of 0.88. The model can assist physicians in small bowel polyp detection during wireless capsule endoscopy diagnosis.
Collecting a sufficient amount of annotated data for training deep learning models is challenging in certain applications, especially medical ones. Our proposed procedure provides a data augmentation method for tasks with extremely small target-domain datasets, enhancing their training data. Consequently, this procedure holds the potential to develop segmentation models in other tasks with small datasets.

[1] Jorge Agnese, Jonathan Herrera, Haicheng Tao, and Xingquan Zhu. A survey and taxonomy of adversarial neural networks for text-to-image synthesis. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(4):e1345, 2020.
[2] Haya Alaskar, Abir Hussain, Nourah Al-Aseem, Panos Liatsis, and Dhiya Al-Jumeily. Application of convolutional neural networks for automated ulcer detection in wireless capsule endoscopy images. Sensors, 19(6):1265, 2019.
[3] Haneen Alokasi and Muhammad Bilal Ahmad. The accuracy performance of semantic segmentation network with different backbones. In 2022 7th International Conference on Data Science and Machine Learning Applications (CDMA), pages 49–54, 2022.
[4] Tomonori Aoki, Atsuo Yamada, Kazuharu Aoyama, Hiroaki Saito, Akiyoshi Tsuboi, Ayako Nakada, Ryota Niikura, Mitsuhiro Fujishiro, Shiro Oka, Soichiro Ishihara, et al. Automatic detection of erosions and ulcerations in wireless capsule endoscopy images based on a deep convolutional neural network. Gastrointestinal endoscopy, 89(2):357–363, 2019.
[5] Halim Awadie, Amir Klein, David Tate, Bilel Jideh, Iddo Bar-Yishai, Kathleen Goodrick, Golo Ahlenstiel, and Michael J. Bourke. The prevalence of small-bowel polyps on video capsule endoscopy in patients with sporadic duodenal or ampullary adenomas. Gastrointestinal Endoscopy, 93(3):630–636, 2021.
[6] Sagie Benaim and Lior Wolf. One-sided unsupervised domain mapping. Advances in neural information processing systems, 30, 2017.
[7] Siddharth Bhatia, Yash Sinha, and Lavika Goel. Lung cancer detection: a deep learning approach. In Soft Computing for Problem Solving: SocProS 2017, Volume 2, pages 699–705. Springer, 2019.
[8] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255, 2009.
[9] Robert A Enns, Lawrence Hookey, David Armstrong, Charles N Bernstein, Steven J Heitman, Christopher Teshima, Grigorios I Leontiadis, Frances Tse, and Daniel Sadowski. Clinical practice guidelines for the use of video capsule endoscopy. Gastroenterology, 152(3):497–514, 2017.
[10] Huan Fu, Mingming Gong, Chaohui Wang, Kayhan Batmanghelich, Kun Zhang, and Dacheng Tao. Geometry-consistent generative adversarial networks for one-sided unsupervised domain mapping. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2427–2436, 2019.
[11] Sirichart Gobpradit and Peerapon Vateekul. Angiodysplasia segmentation on capsule endoscopy images using albunet with squeeze-and-excitation blocks. In Intelligent Information and Database Systems: 12th Asian Conference, ACIIDS 2020, Phuket, Thailand, March 23–26, 2020, Proceedings, Part I 12, pages 283–293. Springer, 2020.
[12] Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Commun. ACM, 6 (11):139–144, oct 2020.
[13] Samir Gupta, David Lieberman, Joseph C Anderson, Carol A Burke, Jason A Dominitz, Tonya Kaltenbach, Douglas J Robertson, Aasma Shaukat, Sapna Syngal, and Douglas K Rex. Recommendations for follow-up after colonoscopy and polypectomy: a consensus update by the us multi-society task force on colorectal cancer. Gastrointestinal endoscopy, 91(3):463–485, 2020.
[14] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, 2016.
[15] Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. Image-to-image translation with conditional adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1125–1134, 2017.
[16] Debesh Jha, Pia H. Smedsrud, Michael A. Riegler, P˚al Halvorsen, Thomas de Lange, Dag Johansen, and H˚avard D. Johansen. Kvasir-seg: A segmented polyp dataset. In Yong Man Ro, Wen-Huang Cheng, Junmo Kim, Wei-Ta Chu, Peng Cui, Jung-Woo Choi, Min-Chun Hu, and Wesley De Neve, editors, MultiMedia Modeling, pages 451–462, Cham, 2020. Springer International Publishing.
[17] Debesh Jha, Nikhil Kumar Tomar, Sharib Ali, Michael A. Riegler, H˚avard D. Johansen, Dag Johansen, Thomas de Lange, and P˚al Halvorsen. Nanonet: Real-time polyp segmentation in video capsule endoscopy and colonoscopy. In 2021 IEEE 34th International Symposium on Computer-Based Medical Systems (CBMS), pages 37–43, 2021.
[18] Alexandros Karargyris and Nikolaos Bourbakis. Detection of small bowel polyps and ulcers in wireless capsule endoscopy videos. IEEE Transactions on Biomedical Engineering, 58(10):2777–2786, 2011.
[19] Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. IEEE Trans. Pattern Anal. Mach. Intell., 43(12):4217–4228, dec 2021.
[20] Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization, 2017.
[21] Bence K˝ov´ari, Baek H Kim, and Gregory Y Lauwers. The pathology of gastric and duodenal polyps: Current concepts. Histopathology, 78(1):106–124, 2021.
[22] Christian Ledig, Lucas Theis, Ferenc Husz´ar, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, et al. Photo-realistic single image super-resolution using a generative adversarial network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4681–4690, 2017.
[23] Hsin-Ying Lee, Hung-Yu Tseng, Jia-Bin Huang, Maneesh Singh, and Ming-Hsuan Yang. Diverse image-to-image translation via disentangled representations. In Proceedings of the European conference on computer vision (ECCV), pages 35–51, 2018.
[24] Ming-Yu Liu, Thomas Breuel, and Jan Kautz. Unsupervised image-to-image translation networks. Advances in neural information processing systems, 30, 2017.
[25] Omid Haji Maghsoudi. Superpixel based segmentation and classification of polyps in wireless capsule endoscopy. In 2017 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), pages 1–4, 2017.
[26] Mehdi Mirza and Simon Osindero. Conditional generative adversarial nets, 2014.
[27] Yingxue Pang, Jianxin Lin, Tao Qin, and Zhibo Chen. Image-to-image translation: Methods and applications. IEEE Transactions on Multimedia, 24:3859–3881, 2021.
[28] Su Bum Park. Application of double-balloon enteroscopy for small bowel tumors. Clinical Endoscopy, 2023.
[29] Taesung Park, Alexei A. Efros, Richard Zhang, and Jun-Yan Zhu. Contrastive learning for unpaired image-to-image translation. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vision – ECCV 2020, pages 319–345, Cham, 2020. Springer International Publishing.
[30] Deepak Pathak, Philipp Krahenbuhl, Jeff Donahue, Trevor Darrell, and Alexei A Efros. Context encoders: Feature learning by inpainting. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2536–2544, 2016.
[31] Marco Pennazio, Emanuele Rondonotti, Edward J Despott, Xavier Dray, Martin Keuchel, Tom Moreels, David S Sanders, Cristiano Spada, Cristina Carretero, Pablo Cortegoso Valdivia, et al. Small-bowel capsule endoscopy and device-assisted enteroscopy for diagnosis and treatment of small-bowel disorders: European society of gastrointestinal endoscopy (esge) guideline–update 2022. Endoscopy, 55(01):58–95, 2023.
[32] Patrick P´erez, Michel Gangnet, and Andrew Blake. Poisson image editing. ACM Trans. Graph., 22(3):313–318, jul 2003.
[33] Zhen Qin, Qingliang Zeng, Yixin Zong, and Fan Xu. Image inpainting based on deep learning: A review. Displays, 69:102028, 2021.
[34] Scott Reed, Zeynep Akata, Xinchen Yan, Lajanugen Logeswaran, Bernt Schiele, and Honglak Lee. Generative adversarial text to image synthesis. In International conference on machine learning, pages 1060–1069. PMLR, 2016.
[35] Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In Nassir Navab, Joachim Hornegger, William M. Wells, and Alejandro F. Frangi, editors, Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pages 234–241, Cham, 2015. Springer International Publishing.
[36] Furqan Rustam, Muhammad Abubakar Siddique, Hafeez Ur Rehman Siddiqui, Saleem Ullah, Arif Mehmood, Imran Ashraf, and Gyu Sang Choi. Wireless capsule endoscopy bleeding images classification using cnn based model. IEEE Access, 9:33675–33688, 2021.
[37] Fatih Saygili, Saba Mukaddes Saygili, and Erkin Oztas. Examining the whole bowel, double balloon enteroscopy: Indications, diagnostic yield and complications. World Journal of Gastrointestinal Endoscopy, 7(3):247, 2015.
[38] Nahian Siddique, Sidike Paheding, Colin P Elkin, and Vijay Devabhaktuni. Unet and its variants for medical image segmentation: A review of theory and applications. Ieee Access, 9:82031–82057, 2021.
[39] Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition, 2015.
[40] Pia H Smedsrud, Vajira Thambawita, Steven A Hicks, Henrik Gjestang, Oda Olsen Nedrejord, Espen Næss, Hanna Borgli, Debesh Jha, Tor Jan Derek Berstad, Sigrun L Eskeland, et al. Kvasir-capsule, a video capsule endoscopy dataset. Scientific Data, 8(1):142, 2021.
[41] Pia H. Smedsrud, Vajira Thambawita, Steven A. Hicks, Henrik Gjestang, Oda Olsen Nedrejord, Espen Næss, Hanna Borgli, Debesh Jha, Tor Jan Derek Berstad, Sigrun L. Eskeland, Mathias Lux, H˚avard Espeland, Andreas Petlund, Duc Tien Dang Nguyen, Enrique Garcia-Ceja, Dag Johansen, Peter T. Schmidt, Ervin Toth, Hugo L. Hammer, Thomas de Lange, Michael A. Riegler, and P˚al Halvorsen. Kvasir-capsule, a video capsule endoscopy dataset. Scientific Data, 8(1):142, May 2021.
[42] Christian Szegedy, Sergey Ioffe, Vincent Vanhoucke, and Alexander A. Alemi. Inception-v4, inception-resnet and the impact of residual connections on learning. In Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence, AAAI’17, page 4278–4284. AAAI Press, 2017.
[43] Luisa F. S´anchez-Peralta, Luis Bote-Curiel, Artzai Pic´on, Francisco M. S´anchez-Margallo, and J. Blas Pagador. Deep learning to find colorectal polyps in colonoscopy: A systematic literature review. Artificial Intelligence in Medicine, 108:101923, 2020.
[44] Chunwei Tian, Xuanyu Zhang, Jerry Chun-Wei Lin, Wangmeng Zuo, Yanning Zhang, and Chia-Wen Lin. Generative adversarial networks for image super-resolution: A survey. arXiv preprint arXiv:2204.13620, 2022.
[45] Anuja Vats, Marius Pedersen, Ahmed Mohammed, and Øistein Hovde. Evaluating clinical diversity and plausibility of synthetic capsule endoscopic images, 2023.
[46] Zhiguo Xiao, Jia Lu, Xiaokun Wang, Nianfeng Li, Yuying Wang, and Nan Zhao. Wce-dcgan: A data augmentation method based on wireless capsule endoscopy images for gastrointestinal disease detection. IET Image Processing, 17(4):1170– 1180, 2023.
[47] Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision, pages 2223–2232, 2017.