Mạng nơ-ron sinh biểu diễn hồi tiếp đa cấp độ và chú ý toàn cục cho việc loại bỏ mưa trong hình ảnh đơn

Neural Computing and Applications - Tập 35 - Trang 3697-3708 - 2022
Meihua Wang1, Chao Li1, Fanhui Ke
1College of Mathematics and Informatics, South China Agricultural University, Guangzhou, China

Tóm tắt

Việc loại bỏ mưa là một bước tiền xử lý thiết yếu cho nhiều nhiệm vụ thị giác máy tính, chẳng hạn như lái xe tự động dựa trên thị giác. Các phương pháp hiện có thường phụ thuộc vào thông tin trước biết hoặc cấu trúc mạng xác định và do đó gặp khó khăn với chi phí tính toán cao. Để nâng cao hiệu suất để đáp ứng yêu cầu thời gian thực của lái xe tự động, chúng tôi đề xuất một mạng nơ-ron hồi tiếp học residual đa cấp độ mới với cơ chế chú ý toàn cục và kiến trúc mạng residual. Trong mạng Hồi tiếp Đa cấp độ Residual và Chú ý Toàn cục (tắt là RMRGN), chúng tôi sử dụng mô hình giai đoạn hồi tiếp để từng bước khai thác thông tin ngữ cảnh toàn cầu và chi tiết hình ảnh nhằm loại bỏ các vệt mưa một cách từ từ. Cơ chế chú ý toàn cục cho phép chúng tôi tập trung vào ý nghĩa ngữ cảnh ở mỗi giai đoạn hồi tiếp, từ đó giúp mạng phân biệt giữa các vệt mưa và hình ảnh không có mưa. Bằng cách khám phá thông tin chú ý, chúng tôi tiến thêm một bước nữa bằng cách đề xuất một mạng học residual sâu đa cấp để loại bỏ các vệt mưa trong một hình ảnh đơn. Các kết quả thực nghiệm toàn diện cho thấy RMRGN hoạt động hiệu quả hơn so với các phương pháp hiện đại nhất trong việc loại bỏ các vệt mưa.

Từ khóa

#mạng nơ-ron hồi tiếp #loại bỏ mưa #học residual #chú ý toàn cục #thị giác máy tính #lái xe tự động

Tài liệu tham khảo

Cao Y, Xu J, Lin S, Wei F, Hu H (2019) Gcnet: Non-local networks meet squeeze-excitation networks and beyond. In: Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, pp 00, Chen YL, Hsu CT (2013) A generalized low-rank appearance model for spatio-temporally correlated rain streaks. In: Proceedings of the IEEE international conference on computer vision, pp 1968–1975 Collins Robert T, Lipton Alan J, Takeo K, Hironobu F, David D, Yanghai T, David T, Nobuyoshi E, Osamu H, Peter B et al (2000) A system for video surveillance and monitoring. VSAM Final Rep 1(2000):1–68 Eigen D, Krishnan D, Fergus R (2013) Restoring an image taken through a window covered with dirt or rain. In: Proceedings of the IEEE international conference on computer vision, pp 633–640 Fan Z, Wu H, Fu X , Hunag Y, Ding X (2018) Residual-guide feature fusion network for single image deraining. arXiv preprint arXiv: 1804.07493 Farenzena M , Bazzani L , Perina A , Murino V , Cristani M (2010) Person re-identification by symmetry-driven accumulation of local features. In: 2010 IEEE computer society conference on computer vision and pattern recognition, pp 2360–2367. IEEE Fu X, Huang J, Ding X, Liao Y, Paisley J (2017) Clearing the skies: a deep network architecture for single-image rain removal. IEEE Trans Image Process 26(6):2944–2956 Fu X, Huang J, Zeng D, Huang Y, Ding X, Paisley J (2017) Removing rain from single images via a deep detail network. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3855–3863 Fu X, Liang B, Huang Y, Ding X, Paisley J (2019) Lightweight pyramid networks for image deraining. IEEE Trans Neural Netw Learn Syst 31(6):1794–1807 Han P, Liu G, Huang S, Yuan W, Lu Z (2016) Segmentation with selectively propagated constraints. In: International conference on neural information processing, pp 585–592. Springer Han P, Shang S, Sun A, Zhao P, Zheng K, Zhang X (2021) Point-of-interest recommendation with global and local context. IEEE Trans Knowl Data Eng He K, Sun J, Tang X (2012) Guided image filtering. IEEE Trans Pattern Anal Mach Intell 35(6):1397–1409 He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 770–778 Hu H, Gu J, Zhang Z, Dai J, Wei Y (2018) Relation networks for object detection. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3588–3597 Huynh-Thu Q, Ghanbari M (2008) Scope of validity of PSNR in image/video quality assessment. Electron Lett 44(13):800–801 Itti L, Koch C, Niebur E (1998) A model of saliency-based visual attention for rapid scene analysis. IEEE Trans Pattern Anal Mach Intell 20(11):1254–1259 Kang LW, Lin CW, Fu YH (2011) Automatic single-image-based rain streaks removal via image decomposition. IEEE Trans Image Process 21(4):1742–1755 Kingma DP, Ba J (2014) A method for stochastic optimization. arXiv:1412.6980 Li X, Wu J, Lin Z, Liu H, Zha H (2018) Recurrent squeeze-and-excitation context aggregation net for single image deraining. In: Proceedings of the European Conference on Computer Vision (ECCV) Li Y, Tan RT, Guo X, Lu J, Brown MS (2016) Rain streak removal using layer priors. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2736–2744 Lin H, Li Y, Fu X, Ding X, Huang Y, Paisley J (2020) Rain o’er me: Synthesizing real rain to derain with data distillation. IEEE Trans Image Process 29:7668–7680 Luo Y, Xu Y, Ji H (2015) Removing rain from a single image via discriminative sparse coding. In: Proceedings of the IEEE International Conference on Computer Vision, pp 3397–3405 Maji S, Berg AC, Malik J (2008) Classification using intersection kernel support vector machines is efficient. In: 2008 IEEE conference on computer vision and pattern recognition, pp 1–8. IEEE Mandic D, Chambers J (2001) Recurrent neural networks for prediction: learning algorithms, architectures and stability. Wiley, New York Mirza M, Osindero S (2014) Conditional generative adversarial nets. arXiv:1411.1784 Nair V, Hinton GE (2010) Rectified linear units improve restricted boltzmann machines. In: Icml Niu Y, Lu Z, Huang S, Han P, Wen JR (2015) Weakly supervised matrix factorization for noisily tagged image parsing. In: IJCAI Paszke A, Gross S, Chintala S, Chanan G, Yang E, DeVito Z, Lin Z, Desmaison A, Antiga L, Lerer A, Automatic differentiation in pytorch, Alban Desmaison Ren D, Zuo W, Hu Q, Zhu P, Meng D (2019) Progressive image deraining networks: A better and simpler baseline. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 3937–3946 Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv: 1409.1556 Sun H, Ang MH, Rus D (2019) A convolutional network for joint deraining and dehazing from a single image for autonomous driving in rain. In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp 962–969. IEEE Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich A (2015) Going deeper with convolutions. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 1–9 Tomasi C, Manduchi R (1998) Bilateral filtering for gray and color images. In: Sixth international conference on computer vision (IEEE Cat. No. 98CH36271), pp 839–846. IEEE Wang X, Girshick R, Gupta A, Kaiming H (2018) Non-local neural networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 7794–7803 Wang Z, Bovik AC, Sheikh HR, Simoncelli EP (2004) Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process 13(4):600–612 Wei W, Meng D, Zhao Q, Xu Z, Wu Y (2019) Semi-supervised transfer learning for image rain removal. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 3877–3886 Xingjian SH, Chen Z, Wang H, Yeung DY, Wong WK, Woo WC (2015) Convolutional lstm network: A machine learning approach for precipitation nowcasting. In: Advances in neural information processing systems, pp 802–810 Yang W, Tan RT, Feng J, Liu J, Guo Z, Yan S (2017) Deep joint rain detection and removal from a single image. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 1357–1366 Yu F, Koltun V (2015) Multi-scale context aggregation by dilated convolutions. arXiv: 1511.07122 Zhang H, Dana K, Shi J, Zhang Z, Wang X, Tyagi A (2018) Context encoding for semantic segmentation. In: Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pp 7151–7160 Zhang H, Sindagi V, Patel VM (2019) Image de-raining using a conditional generative adversarial network. IEEE Trans Circuits Syst Video Technol 30(11):3943–3956 Zhu L, Fu CW, Lischinski D, Heng PA (2017) Joint bi-layer optimization for single-image rain streak removal. In: Proceedings of the IEEE international conference on computer vision, pp 2526–2534