成功大學博碩士論文系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	杜少宏 Tu, Shao-Hung
論文名稱：	應用於水下影像強化之輕量化卷積神經網路電路設計 A Lightweight CNN-based Circuit for Underwater Image Enhancement
指導教授：	陳培殷 Chen, Pei-Yin
學位類別：	碩士 Master
系所名稱：	電機資訊學院 - 資訊工程學系 Department of Computer Science and Information Engineering
論文出版年：	2025
畢業學年度：	113
語文別：	中文
論文頁數：	66
中文關鍵詞：	水下影像、VLSI 、影像增強、卷積神經網路
外文關鍵詞：	Underwater image, VLSI, image enhancement, convolutional neural network
相關次數：	點閱：21 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

水下影像因光的吸收與散射作用導致色偏與對比度降低等問題，造成影像退化、品質不佳。近年來，有許多關於水下影像增強的研究文獻，大致可以分為基於深度學習的方法與非深度學習的方法。深度學習方法能有效提升影像品質，然而多數深度學習的方法架構複雜、參數量龐大，導致硬體實現困難，而非深度學習的方法則是在多變的水下環境往往無法取得穩定的影像增強品質。
本論文提出結合傳統影像增強演算法與輕量化卷積神經網路之水下影像增強模型架構。透過色彩校正、對比度增強等規則式演算法模組輔助網路特徵學習，有效降低模型參數量。此外，本論文亦將此網路架構於硬體實現。
實驗結果顯示，提出的方法在多變的水下影像場景中展現出穩定的增強效果，能避免嚴重失真的情況發生。儘管整體架構所需的參數量遠低於主流深度學習方法，所產出的影像品質在主觀視覺效果與客觀指標的評估仍不遜色於近期相關研究成果。
本論文提出的硬體架構使用Verilog HDL實現，Cell Library製程為TSMC 40nm，使用新思科技的Design Compiler合成。工作頻率可達200MHz ，電路面積為 208779 (〖μm〗^2)，總功率消耗為 64.1545mW。實驗結果證明，此硬體設計具低成本之優勢，並適合部署於資源受限的平台。

Underwater images often suffer from quality degradation such as color distortion and reduced contrast due to light absorption and scattering effects. In recent years, numerous studies have been proposed for underwater image enhancement, which can be broadly categorized into deep learning-based and non-deep learning-based approaches. While deep learning methods can effectively improve image quality, they are often characterized by complex architectures and large parameter sizes, making hardware implementation challenging. On the other hand, non-deep learning methods typically fail to maintain stable enhancement performance under diverse underwater conditions. We propose a hybrid underwater image enhancement framework that combines traditional enhancement techniques with a lightweight convolutional neural network. The framework leverages modules such as color correction and contrast enhancement to assist in feature learning. Experimental results demonstrate that the proposed method achieves stable enhancement across various underwater scenes and avoids severe distortions. Despite having significantly fewer parameters compared to mainstream deep learning models, the proposed approach achieves competitive performance in both subjective visual quality and objective evaluation metrics. Moreover, the method is hardware-friendly and suitable for deployment on resource-constrained platforms.
The proposed hardware architecture is implemented using Verilog HDL and synthesized with TSMC 40nm cell library via Synopsys Design Compiler. For input images of size 256×256, the design operates at a clock frequency of 200 MHz, with a total circuit area of 208,779 μm² and a power consumption of 64.1545 mW.

中文摘要	I
英文摘要	II
誌謝	IX
Contents	X
Table	XIII
Figure Captions	XIV
Chapter 1.	Introduction	1
1 Background	1
1.1 Non-deep Learning based Method	1
1.2 Deep Learning based Method	2
2 Motivation	3
3 Organization	3
Chapter 2.	Related Work	4
1 Improved Retinex[8]	4
1.1 Adaptive Color Correction	4
1.2 Improved Retinex Algorithm	5
2 LANet[9]	6
3 U-Shape[11]	7
4 CCL-Net[13]	8
Chapter 3.	Proposed Method	11
1 Overall Architecture	11
2 LRB	12
2.1 Attention Module	13
3 IEM	15
3.1 ACC	15
3.2 SDCS	17
3.3 YCE	18
4 Loss Function	19
Chapter 4.	Hardware Architecture	21
1 Overall Hardware Architecture	21
2 Memory and Lookup Tables	22
3 RLMDA	22
3.1 Quantization Module and Requantization Module	22
3.1.1 Quantization Design Motivation	22
3.1.2 Quantized Convolution Flow	23
3.1.3 Dyadic Quantization	24
3.1.4 Quantization Module Implementation	24
3.1.5 Requantization Module Implementation	25
3.2 PE Array	25
3.2.1 PE	26
3.2.2 PE Array Data Flow	26
3.3 Other Supporting Modules	28
4 IEM	28
4.1 ACC module	28
4.1.1 Avg Module	28
4.1.2 Log Approximation Module	29
4.1.3 Channel Compensation Module	29
4.2 SDCS&YCE Module	29
4.2.1 RGB2Y Module	30
4.2.2 Clamp Normalization Module	30
4.2.3 Std Module	31
Chapter 5.	Experimental Results	32
1 Experimental Setting	32
2 Ablation Study	33
3 Result	37
3.1 Metric Evaluation	38
3.2 Visual Evaluation	39
4 Hardware Result	43
Chapter 6.	Conclusion	46
References	47
                                    

[1] Kaiming He, Jian Sun and Xiaoou Tang, "Single image haze removal using dark channel prior," 2009 IEEE Conference on Computer Vision and Pattern Recognition,Miami, FL, 2009, pp. 1956-1963.
[2] P. L. J. Drews, E. R. Nascimento, S. S. C. Botelho and M. F. Montenegro Campos, "Underwater Depth Estimation and Image Restoration Based on Single Images," in IEEE Computer Graphics and Applications, vol. 36, no. 2, pp. 24-35, Mar.-Apr. 2016 .
[3] X. Fu, P. Zhuang, Y. Huang, Y. Liao, X. -P. Zhang and X. Ding, "A retinex-based enhancing approach for single underwater image," 2014 IEEE International Conference on Image Processing (ICIP), Paris, France, 2014, pp. 4572-4576.
[4] J. Yuan, W. Cao, Z. Cai and B. Su, "An Underwater Image Vision Enhancement Algorithm Based on Contour Bougie Morphology," in IEEE Transactions on Geoscience and Remote Sensing, vol. 59, no. 10, pp. 8117-8128, Oct. 2021.
[5] M. Yin, X. Du, W. Liu, L. Yu and Y. Xing, "Multiscale Fusion Algorithm for Underwater Image Enhancement Based on Color Preservation," in IEEE Sensors Journal, vol. 23, no. 7, pp. 7728-7740, 1 April1, 2023.
[6] Y. Wang, J. Zhang, Y. Cao and Z. Wang, "A deep CNN method for underwater image enhancement," 2017 IEEE International Conference on Image Processing (ICIP), Beijing, China, 2017, pp. 1382-1386.
[7] Cong, R., Yang, W., Zhang, W., Li, C., Guo, C. L., Huang, Q., & Kwong, S. (2023). Pugan: Physical model-guided underwater image enhancement using gan with dualdiscriminators. IEEE Transactions on Image Processing, 32, 4472-4485.
[8] Lin, S., Li, Z., Zheng, F., Zhao, Q., & Li, S. (2023). Underwater image enhancement based on adaptive color correction and improved retinex algorithm. IEEE Access, 11, 27620-27630.
[9] S. Liu, H. Fan, S. Lin, Q. Wang, N. Ding and Y. Tang, "Adaptive Learning Attention Network for Underwater Image Enhancement," in IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 5326-5333, April 2022.
[10] Johnson, J., Alahi, A., Fei-Fei, L. (2016). Perceptual Losses for Real-Time Style Transfer and Super-Resolution. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds) Computer Vision – ECCV 2016. ECCV 2016.
[11] L. Peng, C. Zhu and L. Bian, "U-Shape Transformer for Underwater Image Enhancement," in IEEE Transactions on Image Processing, vol. 32, pp. 3066-3079, 2023.
[12] Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation." Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18. Springer international publishing, 2015.
[13] Liu, Yi, et al. "Underwater Image Enhancement with Cascaded Contrastive Learning." IEEE Transactions on Multimedia (2024).
[14] Jung-Hao Tsao, "A Special Convolutional Neural Network for Underwater Image Enhancement", unpublished M.A. thesis, National Cheng Kung University, R.O.C.2024.
[15] Y. -H. Chen, T. -J. Yang, J. Emer and V. Sze, "Eyeriss v2: A Flexible Accelerator for Emerging Deep Neural Networks on Mobile Devices," in IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 9, no. 2, pp. 292-308, June 2019.
[16]https://oldcrows.org.tw/download/1131129%20%E4%BA%BA%E5%B7%A5%E6%99%BA%E6%85%A7%E7%A0%94%E8%A8%8E%E6%9C%83_%E7%B0%A1%E6%AD%B7%EF%BC%86%E6%91%98%E8%A6%81%E6%A0%BC%E5%BC%8F_%E8%AC%9B%E8%80%85%E9%99%B3%E5%9D%A4%E5%BF%97%20v2.pdf
[17] K. Panetta, C. Gao and S. Agaian, "Human-Visual-System-Inspired Underwater Image Quality Measures," in IEEE Journal of Oceanic Engineering, vol. 41, no. 3, pp. 541-551, July 2016. Available : https://github.com/xahidbuffon/FUnIE-GAN
[18] M. Yang and A. Sowmya, "An Underwater Color Image Quality Evaluation Metric," in IEEE Transactions on Image Processing, vol. 24, no. 12, pp. 6062-6071, Dec. 2015. Available : https://github.com/bilityniu/underimage-fusion-enhancement
[19] Mittal, A., R. Soundararajan, and A. C. Bovik. "Making a Completely Blind Image Quality Analyzer." IEEE Signal Processing Letters. Vol. 22, Number 3, March 2013,pp. 209–212.
[20] C. -Y. Li, J. -C. Guo, R. -M. Cong, Y. -W. Pang and B. Wang, "Underwater Image Enhancement by Dehazing With Minimum Information Loss and Histogram Distribution Prior," in IEEE Transactions on Image Processing, vol. 25, no. 12, pp.5664-5677, Dec. 2016.
[21] C. Li et al., "An Underwater Image Enhancement Benchmark Dataset and Beyond," in IEEE Transactions on Image Processing, vol. 29, pp. 4376-4389, 2020

校內：2030-07-16公開
校外：2030-07-16公開電子論文尚未授權公開，紙本請查館藏目錄

簡易檢索 / 詳目顯示

相關論文