Affiliation:
1. Wuhan University of Technology
2. Huazhong University of Science and Technology
Abstract
Abstract
Techniques for removing image reflections can effectively eliminate artifacts superimposed on the subject due to factors such as reflections in light or through glass. Most current methods are designed on the assumption that reflection areas maintain the original image content. However, the neglect some extreme cases where there is no original information left, such as photos taken in museums with generated light spots.In this paper, we propose a novel model capable of further removing light spots reflections. Specifically, it takes an image with reflection contamination as input, and then guided by the proposed reflection classifiers and structure restorer, ultimately outputs a predicted transmission layer image.Experimental results demonstrate that the proposed model is applicable to different categories of reflection images, outperforming state-of-the-art reflection removal techniques.In summary, the proposed model improves the effect of image reflection technology based on artificial intelligence(AI) in the case of spot reflection.
Publisher
Research Square Platform LLC
Reference38 articles.
1. Fortunato, S. (2010) Community detection in graphs. Phys. Rep.-Rev. Sec. Phys. Lett. 486: 75-174
2. Yoav Y. Schechner and Nahum Kiryati and Ronen Basri (2000) Separation of Transparent Layers using Focus. Int. J. Comput. Vis. 39: 25--39
3. Anat Levin and Assaf Zomet and Yair Weiss (2002) Learning to Perceive Transparency from the Statistics of Natural Scenes. 1247--1254, Advances in Neural Information Processing Systems 15 [Neural Information Processing Systems, {NIPS} 2002, December 9-14, 2002, Vancouver, British Columbia, Canada]
4. Anat Levin and Assaf Zomet and Yair Weiss (2004) Separating Reflections from a Single Image Using Local Features. 306--313, 2004 {IEEE} Computer Society Conference on Computer Vision and Pattern Recognition {(CVPR} 2004), with CD-ROM, 27 June - 2 July 2004, Washington, DC, {USA}
5. Bernard Sarel and Michal Irani (2004) Separating Transparent Layers through Layer Information Exchange. 328--341, 3024, Lecture Notes in Computer Science, Computer Vision - {ECCV} 2004, 8th European Conference on Computer Vision, Prague, Czech Republic, May 11-14, 2004. Proceedings, Part {IV}, Tom{\'{a}}s Pajdla and Jiri Matas