近年來，肺癌在全球出現的比例極高，人們開始對肺有關注，許多專家學者開始對肺癌進行相關的研究，肺癌的前身是肺結節，我們能藉由肺結節去探討演化成肺癌的可能性，肺結節主要分成四大類，實心結節、鈣化結節、毛玻璃病變結節與部分實心結節，判斷出位置與分辨出肺結節能夠使病患提早發現並治療。
本實驗使用YOLO技術進行分類，並使用進期最新的版本YOLOv8，在架構中進行修改，加入了兩種注意力機制，Convolutional Block Attention Module (CBAM)與Polarized Self-Attention (PSA)，CBAM加入骨幹中，PSA加入頸部中，能讓YOLOv8技術透過注意力機制去加強CT影像特徵提取，將影像輸出後準確去判斷肺結節的種類與位置。

In recent years, the global incidence of lung cancer has risen sharply, prompting increased awareness about lung health. Experts and scholars have focused on researching lung cancer, which often originates from lung nodules. By studying lung nodules, we can explore the possibility of their evolution into lung cancer. Lung nodules are mainly classified into four categories: solid nodules, calcified nodules, ground glass opacity nodule, and part-solid nodules. Identifying their location and distinguishing the types can help patients detect and treat lung cancer early.
This study utilizes YOLO technology for classification, specifically the latest version, YOLOv8, with modifications of adding two attention mechanisms: the Convolutional Block Attention Module (CBAM) in the Backbone and Polarized Self-Attention (PSA) in the Neck. These additions empower YOLOv8 technology to boost the feature extraction of CT images, allowing precise identification of the type and location of lung nodules from the image outputs.  

目錄
第一章 緒論	1
1.1研究動機	1
1.2研究目的	3
1.3 論文組織與架構	4
第二章 相關研究介紹	5
2.1肺結節	5
2.2電腦斷層影像	9
2.3 YOLO	11
2.4 注意力機制	13
2.4.1 Convolutional Block Attention Module (CBAM)	15
2.4.2 Polarized Self-Attention (PSA)	17
2.5 方法相關探討	19
第三章 研究方法	20
3.1 YOLOv8架構圖	20
3.2 YOCP 架構圖	26
3.3 YOCP Backbone	28
3.4 YOCP Neck	33
3.5 YOCP流程圖	36
第四章 實驗結果與討論	37
4.1資料來源	37
4.2 評估指標	39
4.3實驗結果	41
4.4 模型呈現數據	44
第五章 結論與未來展望	47
5.1 結論	47
5.2未來展望	47
參考文獻	49

圖目錄
圖1	 無肺結節（健康人）	7
圖2	 四大種類肺結節，(a) SN、(b) CN、(c) GGN、和(d) PSN	8
圖3	 C3模塊圖	21
圖4	 C2f模塊圖	22
圖5	 Bottleneck模塊圖	24
圖6	 SPPF模塊圖	24
圖7	 YOLOv8架構圖	25
圖8	 YOCP架構圖	27
圖9	 PSA架構圖(parallel)	30
圖10	 PSA架構圖(sequential)	31
圖11	CBAM架構圖	33
圖12	CBAM架構圖(channel)	34
圖13	CBAM架構圖(spatial)	34
圖14	YOCP流程圖	36
圖 15	YOCP模型訓練圖	38
圖16	模型數據分析	42
圖17	YOCP模型各數據	42
圖18	YOCP模型F1數據	43
圖19	偵測肺結節(SN)	44
圖20	偵測肺結節(CN)	45
圖21	偵測肺結節(GGN)	45
圖22	偵測肺結節(PSN)	46
圖23	偵測肺結節（健康人，無肺結節）	46

表目錄
表 1  實驗硬體版本	38
表 2  實驗軟體版本	38
表 3  混淆矩陣	39
表 4  模型數據比較	41 

參考文獻
[1].	康健. 十大癌症榜首換人！肺癌發生人數奪冠，如何早期發現？(2023). 2023  [cited 2024 June 02]; Available from: https://www.commonhealth.com.tw/article/88402.
[2].	健康醫療網. 世界肺癌大會免疫治療研究曝光：存活率有望由2%進步到二位數(2023).2023[cited2024June02];Availablefrom:https://www.healthnews.com.tw/article/59419/.
[3].	Binson, V. and M. Subramoniam. Advances in early lung cancer detection: A systematic review. in 2018 International Conference on Circuits and Systems in Digital Enterprise Technology (ICCSDET). 2018. IEEE.
[4].	Ost, D.E. and M.K. Gould, Decision making in patients with pulmonary nodules. American journal of respiratory and critical care medicine, 2012. 185(4): p. 363-372.
[5].	Jiang, P., et al., A Review of Yolo algorithm developments. Procedia computer science, 2022. 199: p. 1066-1073.
[6].	Monkam, P., et al., Detection and classification of pulmonary nodules using convolutional neural networks: a survey. Ieee Access, 2019. 7: p. 78075-78091.
[7].	Knudsen, E.I., Fundamental components of attention. Annu. Rev. Neurosci., 2007. 30: p. 57-78.
[8].	Woo, S., et al. Cbam: Convolutional block attention module. in Proceedings of the European conference on computer vision (ECCV). 2018.
[9].	Liu, H., et al., Polarized self-attention: Towards high-quality pixel-wise regression. arXiv preprint arXiv:2107.00782, 2021.
[10].	Reeves, A.P., et al., On measuring the change in size of pulmonary nodules. IEEE transactions on medical imaging, 2006. 25(4): p. 435-450.
[11].	Mori, M., et al., Atypical adenomatous hyperplasia of the lung: a probable forerunner in the development of adenocarcinoma of the lung. Modern Pathology, 2001. 14(2): p. 72-84.
[12].	Way, T.W., et al., Computer‐aided diagnosis of pulmonary nodules on CT scans: improvement of classification performance with nodule surface features. Medical physics, 2009. 36(7): p. 3086-3098.
[13].	Zhao, J., et al., Covid-ct-dataset: a ct scan dataset about covid-19. 2020.
[14].	Tuder, R.M. and I. Petrache, Pathogenesis of chronic obstructive pulmonary disease. The Journal of clinical investigation, 2012. 122(8): p. 2749-2755.
[15].	Saltzman, C.L., et al., Coronal plane rotation of the first metatarsal. Foot & ankle international, 1996. 17(3): p. 157-161.
[16].	Redmon, J., et al. You only look once: Unified, real-time object detection. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
[17].	Mijwil, M.M., et al., The Distinction between R-CNN and Fast RCNN in Image Analysis: A Performance Comparison. Asian Journal of Applied Sciences, 2022. 10(5).
[18].	Kesana, A., et al. Brain Tumor Detection Using YOLOv5 and Faster R-CNN. in 2023 2nd International Conference on Vision Towards Emerging Trends in Communication and Networking Technologies (ViTECoN). 2023. IEEE.
[19].	Hu, M., et al. A2-FPN: Attention aggregation based feature pyramid network for instance segmentation. in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021.
[20].	张宸嘉, 朱磊, and 俞璐, 卷积神经网络中的注意力机制综述. Journal of Computer Engineering & Applications, 2021. 57(20).
[21].	Bahdanau, D., K. Cho, and Y. Bengio, Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473, 2014.
[22].	Luong, M.-T., H. Pham, and C.D. Manning, Effective approaches to attention-based neural machine translation. arXiv preprint arXiv:1508.04025, 2015.
[23].	Vaswani, A., et al., Attention is all you need. Advances in neural information processing systems, 2017. 30.
[24].	Bello, I., et al. Attention augmented convolutional networks. in Proceedings of the IEEE/CVF international conference on computer vision. 2019.
[25].	Mei, J., et al., SANet: A slice-aware network for pulmonary nodule detection. IEEE transactions on pattern analysis and machine intelligence, 2021. 44(8): p. 4374-4387.
[26].	Kirillov, A., et al. Segment anything. in Proceedings of the IEEE/CVF International Conference on Computer Vision. 2023.