請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/97934完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 傅楸善 | zh_TW |
| dc.contributor.advisor | Chiou-Shann Fuh | en |
| dc.contributor.author | 陳婷 | zh_TW |
| dc.contributor.author | Ting Chen | en |
| dc.date.accessioned | 2025-07-23T16:09:35Z | - |
| dc.date.available | 2025-07-24 | - |
| dc.date.copyright | 2025-07-23 | - |
| dc.date.issued | 2025 | - |
| dc.date.submitted | 2025-07-08 | - |
| dc.identifier.citation | [1] N. Ahn, B. Kang, and K.-A. Sohn, "Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network," Proceedings of European Conference on Computer Vision, Munich, Germany, pp. 256-272, 2018.
[2] H. Chen, X. Lu, S. Li et al., "Improving Aluminum Surface Defect Super-Resolution with Diffusion Models and Skip Connections," Materials Today Communications, vol. 42, no. 111297, pp. 1-9, doi: https://doi.org/10.1016/j.mtcomm.2024.111297, 2025. [3] H. Chen, Y. Wang, T. Guo et al., "Pre-Trained Image Processing Transformer," doi: 10.48550/arXiv.2012.00364, 2020. [4] H. Chen, X. He, L. Qing et al., "Real-World Single Image Super-Resolution: A Brief Review," Information Fusion, vol. 79, pp. 124-145, doi: https://doi.org/10.1016/j.inffus.2021.09.005, 2022. [5] X. Chen, X. Wang, W. Zhang et al., "HAT: Hybrid Attention Transformer for Image Restoration," doi: 10.48550/arXiv.2309.05239, 2023. [6] S.-C. Chu, Z.-C. Dou, J.-S. Pan et al., "HMANet: Hybrid Multi-Axis Aggregation Network for Image Super-Resolution," doi: 10.48550/arXiv.2405.05001, 2024. [7] T. Dai, J. Cai, Y. Zhang et al., "Second-Order Attention Network for Single Image Super-Resolution," Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA, pp. 11057-11066, doi: 10.1109/CVPR.2019.01132, 2019. [8] C. Dong, C. C. G. Loy, K. M. He et al., "Image Super-Resolution Using Deep Convolutional Networks," doi: 10.48550/arXiv.1501.00092, 2014. [9] C. Dong, C. C. G. Loy, K. M. He et al., "Learning a Deep Convolutional Network for Image Super-Resolution," Proceedings of European Conference on Computer Vision, Zurich, Switzerland, pp. 184-199, 2014. [10] S. Elfwing, E. Uchibe, and K. Doya, "Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning," doi: 10.48550/arXiv.1702.03118, 2017. [11] K. He, X. Zhang, S. Ren et al., "Deep Residual Learning for Image Recognition," Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, pp. 1-12, doi: 10.1109/cvpr.2016.90, 2016. [12] A. G. Howard, M. Zhu, B. Chen et al., " MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications," doi: 10.48550/arXiv.1704.04861, 2017. [13] C.-C. Hsu, C.-M. Lee, and Y.-S. Chou, "DRCT: Saving Image Super-Resolution Away from Information Bottleneck," doi: 10.48550/arXiv.2404.00722, 2024. [14] Y. Huang, L. Shao, and A. F. Frangi, "Simultaneous Super-Resolution and Cross-Modality Synthesis of 3d Medical Images Using Weakly-Supervised Joint Convolutional Sparse Coding," doi: 10.48550/arXiv.1705.02596, 2017. [15] Z. Hui, X. Gao, Y. Yang et al., "Lightweight Image Super-Resolution with Information Multi-Distillation Network," doi: 10.48550/arXiv.1909.11856, 2019. [16] S. Ioffe and C. Szegedy, "Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift," doi: 10.48550/arXiv.1502.03167, 2015. [17] Y. Jo, S. W. Oh, J. Kang et al., "Deep Video Super-Resolution Network Using Dynamic Upsampling Filters without Explicit Motion Compensation," Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, pp. 3224-3232, doi: 10.1109/CVPR.2018.00340, 2018. [18] J. Kim, J. K. Lee, and K. M. Lee, "Accurate Image Super-Resolution Using Very Deep Convolutional Networks," doi: 10.48550/arXiv.1511.04587, 2015. [19] J. Kim, J. K. Lee, and K. M. Lee, "Deeply-Recursive Convolutional Network for Image Super-Resolution," doi: 10.48550/arXiv.1511.04491, 2015. [20] M. Konnik and J. Welsh, "High-Level Numerical Simulations of Noise in CCD and CMOS Photosensors: Review and Tutorial," doi: 10.48550/arXiv.1412.4031, 2014. [21] W.-S. Lai, J.-B. Huang, N. Ahuja et al., "Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution," doi: 10.48550/arXiv.1704.03915, 2017. [22] C. Ledig, L. Theis, F. Huszar et al., "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network," doi: 10.48550/arXiv.1609.04802, 2016. [23] Z. Li, Y. Liu, X. Chen et al., "Blueprint Separable Residual Network for Efficient Image Super-Resolution," doi: 10.48550/arXiv.2205.05996, 2022. [24] J. Liang, J. Cao, G. Sun et al., "SwinIR: Image Restoration Using Swin Transformer," doi: 10.48550/arXiv.2108.10257, 2021. [25] B. Lim, S. Son, H. Kim et al., "Enhanced Deep Residual Networks for Single Image Super-Resolution," doi: 10.48550/arXiv.1707.02921, 2017. [26] H. Lin, X. Cheng, X. Wu et al., "CAT: Cross Attention in Vision Transformer," doi: 10.48550/arXiv.2106.05786, 2021. [27] X. Luo, Y. Xie, Y. Zhang et al., "Latticenet: Towards Lightweight Image Super-Resolution with Lattice Block," Proceedings of European Conference on Computer Vision, Glasgow, UK, pp. 272-289. 2020. [28] Y. Mei, Y. Fan, and Y. Zhou, "Image Super-Resolution with Non-Local Sparse Attention," Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Nashville, TN, pp. 3516-3525, doi: 10.1109/CVPR46437.2021.00352, 2021. [29] P. Milanfar, Super-Resolution Imaging, 1 ed.: CRC Press, Boca Raton, Florida, https://doi.org/10.1201/9781439819319, 2011. [30] B. Niu, W. Wen, W. Ren et al., "Single Image Super-Resolution Via a Holistic Attention Network," doi: 10.48550/arXiv.2008.08767, 2020. [31] P. Rasti, H. Demirel, and G. Anbarjafari, "Iterative Back Projection Based Image Resolution Enhancement," Proceedings of Conference on Machine Vision and Image Processing, Zanjan, Iran, pp. 237-240, doi: 10.1109/IranianMVIP.2013.6779986, 2013. [32] F. Salvetti, V. Mazzia, A. Khaliq et al., "Multi-Image Super Resolution of Remotely Sensed Images Using Residual Attention Deep Neural Networks," Remote Sensing, vol. 12, pp. 1-20, doi: 10.3390/rs12142207, 2020. [33] M. Sandler, A. Howard, M. Zhu et al., " MobileNetV2: Inverted Residuals and Linear Bottlenecks," doi: 10.48550/arXiv.1801.04381, 2018. [34] W. Shi, J. Caballero, F. Huszár et al., "Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network," doi: 10.48550/arXiv.1609.05158, 2016. [35] Y. Tai, J. Yang, and X. Liu, "Image Super-Resolution Via Deep Recursive Residual Network," Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Hawaii, pp. 2790-2798, doi: 10.1109/CVPR.2017.298, 2017. [36] R. Y. Tsai and T. S. Huang, "Multiframe Image Restoration and Registration," Multiframe Image Restoration and Registration, vol. 1, pp. 317-339, https://ui.adsabs.harvard.edu/abs/1984ACVIP...1..317T, 1984. [37] P. Wang, B. Bayram, and E. Sertel, "A Comprehensive Review on Deep Learning Based Remote Sensing Image Super-Resolution Methods," Earth-Science Reviews, vol. 232, no. 104110, pp. 1-25, doi: https://doi.org/10.1016/j.earscirev.2022.104110, 2022. [38] H. Zhai, H. Li, and W. Gao, "Multi-Frame Image Super Resolution Using Icm Enhancement Method," Information Technology Journal, vol. 13, no. 14, pp. 2277-2283, doi: 10.3923/itj.2014.2277.2283, 2014. [39] J. Zhang, Y. Zhang, J. Gu et al., "Accurate Image Restoration with Attention Retractable Transformer," doi: 10.48550/arXiv.2210.01427, 2022. [40] Q. Zhang, Y. Li, B. Steele et al., "Comparison of a CMOS-Based and a CCD-Based Digital X-Ray Imaging System: Observer Studies," Journal of Electronic Imaging, vol. 14, no. 2, pp. 1-6, https://doi.org/10.1117/1.1902763, 2005. [41] X. Zhang, H. Zeng, S. Guo et al., "Efficient Long-Range Attention Network for Image Super-Resolution," doi: 10.48550/arXiv.2203.06697, 2022. [42] Y. Zhang, K. Li, K. Li et al., "Image Super-Resolution Using Very Deep Residual Channel Attention Networks," doi: 10.48550/arXiv.1807.02758, 2018. [43] Y. Zhou, M. Yuan, J. Zhang et al., "Review of Vision-Based Defect Detection Research and Its Perspectives for Printed Circuit Board," Journal of Manufacturing Systems, vol. 70, pp. 557-578, doi: https://doi.org/10.1016/j.jmsy.2023.08.019, 2023. [44] Test Research, Inc., “TRI Innovation,” https://www.tri.com.tw/tw/index.html, 2023. [45] M. C. Hsu, “HsuSR: Super Resolution for Printed Circuit Board X-Ray Images,” Master Thesis, Graduate Institute of Networking and Multimedia, National Taiwan University, 2023. [46] N. Otsu, "A threshold selection method from gray-level histograms," Automatica, vol. 11, pp. 285-296, 1975. | - |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/97934 | - |
| dc.description.abstract | 隨著電子產品高度精密化，印刷電路板（Printed Circuit Board, PCB）設計日趨複雜，對於瑕疵檢測的精度與效率提出更高要求。X光自動檢測（Automated X-ray Inspection, AXI）雖能穿透元件結構，協助發現隱藏瑕疵，但其所獲得之影像常因硬體限制而解析度不足，進而影響後續瑕疵辨識準確性。因此，本論文針對單張X光影像解析度提升問題，提出一種輕量級深度學習超解析度演算法ChenSR，以提升影像細節還原能力，同時保有高運算效率。
本論文所提出之ChenSR架構為改良式的Transformer模型，結合Efficient Hybrid Transformer Blocks（EHTB）與MobileNetV2中的inverted residual設計，有效降低模型參數量與記憶體需求，並透過通道注意力與多層殘差連接保持重建品質。與其他主流深度學習超解析度方法相比，ChenSR展現出極具競爭力的推論速度，使其成為目前深度學習架構中運算效率表現相當優異的選擇之一，特別適合部署於即時或邊緣裝置之工業應用場景。 在實驗設計上，本研究與現有代表性方法 （Bicubic、HsuSR、DRCT與HMANet） 進行比較。實驗結果顯示，ChenSR在影像品質（PSNR與SSIM）、結構誤差評估（Diff）及執行效率（計算時間）方面皆展現優異之平衡表現，同時在0.03秒內完成單張影像4倍放大，滿足即時檢測應用需求。 | zh_TW |
| dc.description.abstract | With the increasing complexity and miniaturization of electronic products, the design of Printed Circuit Boards (PCBs) has become more intricate, raising higher demands for both the accuracy and efficiency of defect inspection. Although Automated X-ray Inspection (AXI) can penetrate component structures to detect hidden defects, the acquired images often suffer from insufficient resolution due to hardware limitations, which negatively impacts the accuracy of subsequent defect recognition. To address the challenge of enhancing single X-ray image resolution, this thesis proposes a lightweight deep learning-based super-resolution algorithm, ChenSR, which aims to recover fine image details while maintaining high computational efficiency.
Our proposed ChenSR architecture is a modified Transformer-based model, incorporating Efficient Hybrid Transformer Blocks (EHTB) and the inverted residual structure from MobileNetV2. This design effectively reduces model parameters and memory usage while maintaining reconstruction quality through channel attention and multi-level residual connections. Compared with other mainstream deep learning super-resolution methods, ChenSR exhibits highly competitive inference speed, making it one of the most efficient frameworks currently available. Its lightweight and efficient design makes it particularly suitable for real-time or edge-device deployment in industrial applications. In the experimental evaluation, ChenSR is compared with several representative methods, including Bicubic, HsuSR, DRCT, and HMANet. Results demonstrate that ChenSR achieves a strong balance between image quality (measured by PSNR and SSIM), structural error (Diff), and inference time. Notably, it completes 4× image enlargement within 0.03 seconds per image, fulfilling the requirements for real-time defect inspection. | en |
| dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2025-07-23T16:09:35Z No. of bitstreams: 0 | en |
| dc.description.provenance | Made available in DSpace on 2025-07-23T16:09:35Z (GMT). No. of bitstreams: 0 | en |
| dc.description.tableofcontents | 口試委員會審定書 i
誌謝 ii 中文摘要 iii ABSTRACT v CONTENTS vii LIST OF FIGURES x LIST OF TABLES xiii Chapter 1 Introduction 1 1.1 Importance of Resolution in Digital Imaging 1 1.2 Challenges and Constraints in Image Resolution 2 1.3 Super-Resolution Techniques for Image Reconstruction 4 1.3.1 Single-Image Super Resolution (SISR) 5 1.3.2 Multi-Image Super Resolution (MISR) 6 1.4 Thesis Organization 8 Chapter 2 Related Works 9 2.1 CNN-Based Methods for Single-Image Super Resolution 10 2.2 Transformer-Based Methods for Single-Image Super Resolution 13 2.3 Efficient Single-Image Super Resolution Methods 16 Chapter 3 Background 20 3.1 Process of Defect Inspection 22 3.2 Design Strategy for the Super-Resolution Architecture 23 Chapter 4 Methodology 26 4.1 Overview 26 4.2 ChenSR Algorithm 28 4.2.1 Shallow Feature Extraction 28 4.2.2 Deep Feature Extraction 29 4.2.3 Image Reconstruction 31 4.3 Improving Execution Efficiency 33 4.3.1 Incorporation of Lightweight Convolutional Modules 33 4.3.2 Simplification of the DRCT Backbone 37 4.3.3 Strategies for Preserving Reconstruction Quality 40 Chapter 5 Experimental Results 43 5.1 System and Parameters Setting 43 5.2 Evaluation Metrics 44 5.2.1 PSNR (Peak Signal-to-Noise Ratio) 45 5.2.2 SSIM (Structural Similarity Index Measure) 46 5.2.3 Pixel-wise Difference Evaluation (Diff) 47 5.2.4 Computation Time for 4× Super-Resolution Reconstruction 48 5.3 Evaluation Dataset Description 49 5.4 Experimental Results of Image Super-Resolution 54 Chapter 6 Conclusion and Future Works 77 References 79 | - |
| dc.language.iso | en | - |
| dc.subject | 印刷電路板 | zh_TW |
| dc.subject | 超解析度 | zh_TW |
| dc.subject | 深度學習 | zh_TW |
| dc.subject | 單張X光影像 | zh_TW |
| dc.subject | 瑕疵檢測 | zh_TW |
| dc.subject | 印刷電路板 | zh_TW |
| dc.subject | 超解析度 | zh_TW |
| dc.subject | 深度學習 | zh_TW |
| dc.subject | 單張X光影像 | zh_TW |
| dc.subject | 瑕疵檢測 | zh_TW |
| dc.subject | Printed Circuit Boards | en |
| dc.subject | super-resolution | en |
| dc.subject | Printed Circuit Boards | en |
| dc.subject | defect inspection | en |
| dc.subject | single X-ray image | en |
| dc.subject | deep learning | en |
| dc.subject | super-resolution | en |
| dc.subject | deep learning | en |
| dc.subject | single X-ray image | en |
| dc.subject | defect inspection | en |
| dc.title | 陳超解析度：利用超解析度增強印刷電路板X光影像 | zh_TW |
| dc.title | ChenSR: Enhancing Printed Circuit Board X-Ray Images Using Super Resolution | en |
| dc.type | Thesis | - |
| dc.date.schoolyear | 113-2 | - |
| dc.description.degree | 碩士 | - |
| dc.contributor.oralexamcommittee | 巫宗昇;方瓊瑤 | zh_TW |
| dc.contributor.oralexamcommittee | Zong-Sheng Wu;Chiung-Yao Fang | en |
| dc.subject.keyword | 印刷電路板,瑕疵檢測,單張X光影像,深度學習,超解析度, | zh_TW |
| dc.subject.keyword | Printed Circuit Boards,defect inspection,single X-ray image,deep learning,super-resolution, | en |
| dc.relation.page | 85 | - |
| dc.identifier.doi | 10.6342/NTU202501596 | - |
| dc.rights.note | 未授權 | - |
| dc.date.accepted | 2025-07-10 | - |
| dc.contributor.author-college | 電機資訊學院 | - |
| dc.contributor.author-dept | 生醫電子與資訊學研究所 | - |
| dc.date.embargo-lift | N/A | - |
| 顯示於系所單位： | 生醫電子與資訊學研究所 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-113-2.pdf 未授權公開取用 | 2.76 MB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。