Zobrazeno 1 - 10
of 167
pro vyhledávání: '"Sabine Süsstrunk"'
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783031250620
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::d6b41953742c81a92a902eabfcabff5e
https://doi.org/10.1007/978-3-031-25063-7_22
https://doi.org/10.1007/978-3-031-25063-7_22
Publikováno v:
London Imaging Meeting. 2:38-42
Autor:
Di Chang, Aljaž Božič, Tong Zhang, Qingsong Yan, Yingcong Chen, Sabine Süsstrunk, Matthias Nießner
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783031198205
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::4e911482a57dfb2f4c2cc41ea1f355ec
https://doi.org/10.1007/978-3-031-19821-2_38
https://doi.org/10.1007/978-3-031-19821-2_38
Publikováno v:
City University of Hong Kong
Robustness to adversarial attacks was shown to require a larger model capacity, and thus a larger memory footprint. In this paper, we introduce an approach to obtain robust yet compact models by pruning randomly-initialized binary networks. Unlike ad
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::8546ee4153ad72807ca6f51603b26f04
Autor:
Sabine Süsstrunk, Majed El Helou
Publikováno v:
IEEE Transactions on Image Processing. 29:4885-4897
Blind and universal image denoising consists of using a unique model that denoises images with any level of noise. It is especially practical as noise levels do not need to be known when the model is developed or at test time. We propose a theoretica
Publikováno v:
IEEE transactions on neural networks and learning systems.
Training certifiable neural networks enables us to obtain models with robustness guarantees against adversarial attacks. In this work, we introduce a framework to obtain a provable adversarial-free region in the neighborhood of the input data by a po
Autor:
Yuntao Wu, Dongliang He, Fu Li, B. Z. Ding, Sy-Yen Kuo, Zhipeng Luo, Vishal Monga, Majed El Helou, Sabari Nathan, Songhua Liu, Fangya Li, Qing Wang, Maitreya Suin, Tongtong Zhao, Hao-Hsiang Yang, Ming-Ming Cheng, Sabine Süsstrunk, Li Xin, Priya Kansal, Chenghua Li, Zhen Li, Cheng-Ze Lu, Zhongyun Hu, Ntumba Elie Nsampi, Amirsaeed Yazdani, A. N. Rajagopalan, Zuo-Liang Zhu, Shanshan Zhao, Zeng-Sheng Kuang, Wanli Qian, Zhiguang Zhang, Radu Timofte, Ruifeng Deng, Tianwei Lin, Tao Lu, Yuanzhi Wang, Jianye He, Xiu-Li Shao, Wei-Ting Chen, Tiantong Guo, Ruofan Zhou, Yanduo Zhang, Jia-Xiong Qiu, Hao-Lun Luo
Publikováno v:
CVPR Workshops
Image relighting is attracting increasing interest due to its various applications. From a research perspective, im-age relighting can be exploited to conduct both image normalization for domain adaptation, and also for data augmentation. It also has
Publikováno v:
Color and Imaging Conference. 26:1-6
Size uniformity is one of the prominent features of superpixels. However, size uniformity rarely conforms to the varying content of an image. The chosen size of the superpixels therefore represents a compromise - how to obtain the fewest superpixels
Publikováno v:
Computer Vision – ECCV 2020 ISBN: 9783030586034
ECCV (28)
ECCV (28)
Existing techniques to encode spatial invariance within deep convolutional neural networks (CNNs) apply the same warping field to all the feature channels. This does not account for the fact that the individual feature channels can represent differen
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::c456a44d11321db850826e2535e35748
http://arxiv.org/abs/2007.09433
http://arxiv.org/abs/2007.09433
Publikováno v:
ICASSP
Extreme image or video completion, where, for instance, we only retain 1% of pixels in random locations, allows for very cheap sampling in terms of the required pre-processing. The consequence is, however, a reconstruction that is challenging for hum
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::5e3b9cd77b47ecce0b5016fd70328c22
http://arxiv.org/abs/2004.06409
http://arxiv.org/abs/2004.06409