Visionary Fusion: Empowering Robot Sight with Model-driven Multi-band Enhancement
Keywords:
image fusion, image enhancement, topic models, robot vision, autonomous navigationAbstract
Single image enhancement is crucial for improving visual quality in applications like robot vision. This study proposes a novel topic-model assisted multi-band image fusion method to enhance a single input image while preserving semantic information. The key innovation is using topic models for adaptive band fusion based on image content. Comparative experiments on benchmark datasets demonstrate that the proposed technique outperforms state-of-the-art methods in quantitative and qualitative evaluation. This study also provides real-time demonstrations and case studies in robot navigation, highlighting the efficacy of the proposed fusion strategy. The fused images exhibit finer details and color consistency leading to improved navigation and obstacle avoidance. This technique has wide-ranging applications for computer vision related tasks in autonomous systems.
References
Wang Z, Bovik AC, Sheikh HR, Simoncelli EP. Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process. 2004 Apr 13; 13(4): 600–12.
Ma K, Duanmu Z, Wu Q, Wang Z, Yong H, Li H, Zhang L. Waterloo exploration database: New challenges for image quality assessment models. IEEE Trans Image Process. 2016 Nov 22; 26(2): 1004–16.
Fu X, Zeng D, Huang Y, Liao Y, Ding X, Paisley J. A fusion-based enhancing method for weakly illuminated images. Signal Process. 2016 Dec 1; 129: 82–96.
Liu Y, Wang K, Lan H, Lin L. Temporal contrastive graph learning for video action recognition and retrieval. arXiv preprint arXiv:2101.00820. 2021 Jan 4.
Ma K, Wang Z. Multi-exposure image fusion: A patch-wise approach. In 2015 IEEE International Conference on Image Processing (ICIP). 2015 Sep 27; 1717–1721.
Li S, Kang X, Hu J. Image fusion with guided filtering. IEEE Trans Image Process. 2013 Jan 30; 22(7): 2864–75.
Byeon W, Breuel TM, Raue F, Liwicki M. Scene labeling with LSTM recurrent neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition. 2015;
–3555.
Dhamo H, Farshad A, Laina I, Navab N, Hager GD, Tombari F, Rupprecht C. Semantic image manipulation using scene graphs. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020; 5213–5222.
Ren W, Liu S, Zhang H, Pan J, Cao X, Yang MH. Single image dehazing via multi-scale convolutional neural networks. In: Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14. Cham: Springer International Publishing; 2016; 154–169.
Xu H, Ma J, Le Z, Jiang J, Guo X. Fusiondn: A unified densely connected network for image fusion. In Proceedings of the AAAI conference on artificial intelligence. 2020 Apr 3; 34(07): 12484–12491.
Li H, Wu XJ, Kittler J. Infrared and visible image fusion using a deep learning framework. In 2018 IEEE 24th international conference on pattern recognition (ICPR). 2018 Aug 20; 2705–2710.
Zhang Y, Bai X, Wang T. Boundary finding based multi-focus image fusion through multi-scale morphological focus-measure. Inf Fusion. 2017 May 1; 35: 81–101.
Ram Prabhakar K, Sai Srikar V, Venkatesh Babu R. Deepfuse: A deep unsupervised approach for exposure fusion with extreme exposure image pairs. In Proceedings of the IEEE international conference on computer vision. 2017; 4714–4722.
Han XF, Feng ZA, Sun SJ, Xiao GQ. 3D point cloud descriptors: state-of-the-art. Artif Intell Rev. 2023 Apr 12: 56(10): 1–51.
Peng H, Chen Q, Yan M, Cao Z, Li X, Zhang Y, Shi J, Feng GH, Zhang X, Zhang Y, Qian J. Sperm tsRNAs contribute to intergenerational inheritance of an acquired metabolic disorder. Science. 2016 Jan 22; 351(6271): 397–400.
Wang S, Zhang L, Wang X. Single image haze removal via attention-based transmission estimation and classification fusion network. Neurocomputing. 2021 Aug 4; 447: 48–63.
