放射線治療是一種常見的癌症治療方法，使用高能量的放射線來破壞癌細胞的生長和分裂能力。這種治療方法通常用於局部癌症控制，可以用於治療多種癌症，包括頭頸部、乳房、肺部、攝護腺等部位的癌症。放射線治療的過程通常需要一個多週的治療計劃，包括多次的放射線療程。在治療過程中，患者會被安置在治療機器中，放射線從機器中釋放出來，瞄準癌症部位。放射線通過直接擊中癌細胞，破壞其DNA結構，阻止其生長和分裂。同時，正常細胞也會受到放射線的影響，但正常組織通常有較好的修復能力，可以在治療結束後恢復健康。放射線治療的具體劑量和治療計劃根據患者的癌症類型、位置、大小以及個人化的治療目標而定。治療團隊會根據醫學影像學、放射線物理學和腫瘤學等領域的知識，制定一個定制的治療計劃，以最大限度地破壞癌細胞同時最小化對正常組織的損害。放射線治療需要避開重要器官，放射腫瘤學家需要鉤勒出重要器官的位置，這個過程耗時費力。然而，不僅放射線治療需要這樣的器官定位，其他治療方法也面臨相同的挑戰；因此，開發出精確分割特定生理結構的方法將顯著加快治療速度。

這篇論文基於 Attention U-Net 架構，其貢獻可以分為：首先，我們提出了 ABSCBAM，受到 PANet 的啟發，這是在 CBAM 的基礎上新增了一個二元分割預測機制。通過引入輔助的二元分割監督，我們成功提升了模型的性能。相較於其他基準方法，在有限的顯示卡記憶體下，我們僅稍微增加了計算量，卻取得了顯著的性能提升。其次，我們對威斯康辛大學麥迪遜分校提供的胃腸道圖像分割資料集進行了深入分析。為了應對原始資料集的特性，我們將三張影像合併成一張 2.5D 影像，進行更全面的信息提取與處理。最後，我們提供了一系列實驗數據，進行了詳細的實驗來尋找最佳參數設定。這包括卷積核的大小選擇、損失函數的比較等等。這些實驗數據為我們的研究提供了堅實的支持和驗證。

總結而言，本論文不僅在模型架構上提出了創新的 ABSCBAM 機制，更通過深入分析和豐富的實驗數據展示了這一方法的有效性與優越性。我們的研究不僅對圖像分割領域具有實際意義，也為未來相關領域的探索提供了有價值的參考。

Radiation therapy is a common cancer treatment that uses high-energy radiation to destroy the ability of cancer cells to grow and divide. This treatment is often used for localized cancer control and can be used to treat a variety of cancers, including those of the head and neck, breast, lung, prostate, and more.

The course of radiation therapy usually requires a multi-week treatment plan that includes multiple radiation sessions. During treatment, the patient is placed in a treatment machine from which radiation is delivered to target the cancer. Radiation works by directly hitting cancer cells, disrupting their DNA structure, preventing them from growing and dividing. At the same time, normal cells will also be affected by radiation, but normal tissues usually have a better repair ability and can return to health after treatment.

The specific dose and treatment plan for radiation therapy depends on the patient's cancer type, location, size, and individual treatment goals. Drawing on knowledge from the fields of medical imaging, radiation physics, and oncology, the treating team develops a customized treatment plan to maximize the destruction of cancer cells while minimizing damage to normal tissue.

Radiation therapy needs to avoid vital organs, and radiation oncologists need to outline the location of vital organs, which is a time-consuming and laborious process. However, such organ localization is not only required for radiation therapy, but also for other therapeutic modalities; therefore, developing methods to accurately segment specific physiological structures would significantly speed up treatment.

This thesis, based on the Attention U-Net architecture, contributes in several aspects:
Firstly, we propose ABSCBAM, inspired by PANet, which extends CBAM with a binary segmentation prediction mechanism. By introducing auxiliary binary segmentation supervision, we successfully enhance the model's performance. Compared to other baseline methods, we achieve significant performance improvement with only a slight increase in computational demand, all within the constraints of limited GPU memory.Secondly, we conduct an in-depth analysis of the gastrointestinal image segmentation dataset provided by the University of, Wisconsin-Madison. To address the dataset's unique characteristics, we combine three images into a single 2.5D image, enabling more comprehensive information extraction and processing.Lastly, we provide a series of experimental results, meticulously exploring optimal parameter settings. This encompasses choices such as convolutional kernel sizes, as well as comparisons of different loss functions. These empirical findings solidify and validate our research approach.
In summary, this paper not only introduces the innovative ABSCBAM mechanism within the model architecture but also substantiates its effectiveness and superiority through comprehensive analysis and a rich array of experimental data. Our research not only holds practical significance in the field of image segmentation but also offers valuable insights for future explorations in related domains.

[1] Spyridon Bakas, Hamed Akbari, Aristeidis Sotiras, and et al. Advancing the cancer genome atlas glioma mri collections with expert segmentation labels and radiomic features. Scientific Data, 4:170117, 2017.
[2] Laura Mora Ballestar and Veronica Vilaplana. Mri brain tumor segmentation and uncertainty estimation using 3d-unet architectures. In Alessandro Crimi and Spyridon Bakas, editors, Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, pages 376–390, Cham, 2021. Springer International Publishing.
[3] Irwan Bello, Barret Zoph, Quoc Le, Ashish Vaswani, and Jonathon Shlens. Attention augmented convolutional networks. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 3285–3294, 2019.
[4] Alexander Buslaev, Vladimir I. Iglovikov, Eugene Khvedchenya, Alex Parinov, Mikhail Druzhinin, and Alexandr A. Kalinin. Albumentations: Fast and flexible image augmentations. Information, 11(2), 2020.
[5] Özgün Çiçek, Ahmed Abdulkadir, Soeren S. Lienkamp, Thomas Brox, and Olaf Ronneberger. 3d u-net: Learning dense volumetric segmentation from sparse annotation.In Sebastien Ourselin, Leo Joskowicz, Mert R. Sabuncu, Gozde Unal, and William Wells, editors, Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016, pages 424–432, Cham, 2016. Springer International Publishing.
[6] Jun Fu, Jing Liu, Haijie Tian, Yong Li, Yongjun Bao, Zhiwei Fang, and Hanqing Lu. Dual attention network for scene segmentation. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3141–3149, 2019.
[7] Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation networks. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7132–7141, 2018.
[8] Kaggle. UW-Madison GI Tract Image Segmentation Competition. https://www.kaggle.com/competitions/uw-madison-gi-tract-image-segmentation/, 2022.
[9] Geert Litjens, Thijs Kooi, Babak Ehteshami Bejnordi, Arnaud Arindra Adiyoso Setio, Francesco Ciompi, Mohsen Ghafoorian, Jeroen A. W. M. van der Laak, Bram van Ginneken, and Clara I. Sánchez. A survey on deep learning in medical image analysis. CoRR, abs/1702.05747, 2017.
[10] Bjoern H. Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, Justin Kirby, Yuliya Burren, Nicole Porz, Johannes Slotboom, Roland Wiest, Levente Lanczi, Elizabeth Gerstner, Marc-André Weber, Tal Arbel, Brian B. Avants, Nicholas Ayache, Patricia Buendia, D. Louis Collins, Nicolas Cordier, Jason J. Corso, Antonio Criminisi, Tilak Das, Hervé Delingette, Çağatay Demiralp, Christopher R. Durst, Michel Dojat, Senan Doyle, Joana Festa, Florence Forbes, Ezequiel Geremia, Ben Glocker, Polina Golland, Xiaotao Guo, Andac Hamamci, Khan M. Iftekharuddin, Raj Jena, Nigel M. John, Ender Konukoglu, Danial Lashkari, José António Mariz, Raphael Meier, Sérgio Pereira, Doina Precup, Stephen J. Price, Tammy Riklin Raviv, Syed M. S. Reza, Michael Ryan, Duygu Sarikaya, Lawrence Schwartz, Hoo Chang Shin, Jamie Shotton, Carlos A. Silva, Nuno Sousa, Nagesh K. Subbanna, Gabor Szekely, Thomas J. Taylor, Owen M. Thomas, Nicholas J. Tustison, Gozde Unal, Flor Vasseur, Max Wintermark, Dong Hye Ye, Liang Zhao, Binsheng Zhao, Darko Zikic, Marcel Prastawa, Mauricio Reyes, and Koen Van Leemput. The multimodal brain tumor image segmentation benchmark (brats). IEEE Transactions on Medical Imaging, 34(10):1993–2024, 2015.
[11] Ozan Oktay, Jo Schlemper, Loïc Le Folgoc, Matthew C. H. Lee, Mattias P. Heinrich, Kazunari Misawa, Kensaku Mori, Steven G. McDonagh, Nils Y. Hammerla, Bernhard Kainz, Ben Glocker, and Daniel Rueckert. Attention u-net: Learning where to look for the pancreas. CoRR, abs/1804.03999, 2018.
[12] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. Pytorch: An imperative style, high-performance deep learning library. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019.
[13] 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.
[14] Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In So Kweon. Cbam: Convolutional block attention module. In Proceedings of the European Conference on Computer Vision (ECCV), September 2018.
[15] Kang Zhang, Xiaohong Liu, Jun Shen, Zhihuan Li, Ye Sang, Xingwang Wu, Yunfei Zha, Wenhua Liang, Chengdi Wang, Ke Wang, Linsen Ye, Ming Gao, Zhongguo Zhou, Liang Li, Jin Wang, Zehong Yang, Huimin Cai, Jie Xu, Lei Yang, Wenjia Cai, Wenqin Xu, Shaoxu Wu, Wei Zhang, Shanping Jiang, Lianghong Zheng, Xuan Zhang, Li Wang, Liu Lu, Jiaming Li, Haiping Yin, Winston Wang, Oulan Li, Charlotte Zhang, Liang Liang, Tao Wu, Ruiyun Deng, Kang Wei, Yong Zhou, Ting Chen, Johnson Yiu-Nam Lau, Manson Fok, Jianxing He, Tianxin Lin, Weimin Li, and Guangyu Wang. Clinically applicable ai system for accurate diagnosis, quantitative measurements, and prognosis of covid-19 pneumonia using computed tomography. Cell, 181(6):1423–1433.e11, 2020.
[16] Zhengxin Zhang, Qingjie Liu, and Yunhong Wang. Road extraction by deep residual u-net. IEEE Geoscience and Remote Sensing Letters, 15(5):749–753, 2018.
[17] Xiangyu Zhao, Peng Zhang, Fan Song, Chenbin Ma, Guangda Fan, Yangyang Sun, Youdan Feng, and Guanglei Zhang. Prior attention network for multi-lesion segmentation in medical images. IEEE Transactions on Medical Imaging, 41(12):3812–3823, 2022.