Zobrazeno 1 - 10
of 34
pro vyhledávání: '"Zhanghui Kuang"'
Publikováno v:
IEEE Transactions on Image Processing. 30:8306-8317
Human-object interaction detection that aims at detectinghuman, verb, objecttriplets is critical for the holistic human-centric scene understanding. Existing approaches ignore the modeling of correlations among hierarchical human parts and objects. I
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ad6d85f28c25b4b629d61686faef6c6d
https://doi.org/10.1109/tip.2021.3093784
https://doi.org/10.1109/tip.2021.3093784
Transformers with powerful global relation modeling abilities have been introduced to fundamental computer vision tasks recently. As a typical example, the Vision Transformer (ViT) directly applies a pure transformer architecture on image classificat
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::58e96c6bd2033cdc953e2a10a7bd3e33
https://ora.ox.ac.uk/objects/uuid:a222054a-8742-4194-96ec-254bc9d1a85a
https://ora.ox.ac.uk/objects/uuid:a222054a-8742-4194-96ec-254bc9d1a85a
Publikováno v:
Hong Kong University of Science and Technology
Weakly-Supervised Semantic Segmentation (WSSS) segments objects without a heavy burden of dense annotation. While as a price, generated pseudo-masks exist obvious noisy pixels, which result in sub-optimal segmentation models trained over these pseudo
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::dc4c30a1ba44e3593f544ad5b8382091
http://arxiv.org/abs/2112.07431
http://arxiv.org/abs/2112.07431
Publikováno v:
CVPR
One of the main challenges for arbitrary-shaped text detection is to design a good text instance representation that allows networks to learn diverse text geometry variances. Most of existing methods model text instances in image spatial domain via m
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::951c53b1ece2e7a326aa34906161d501
http://arxiv.org/abs/2104.10442
http://arxiv.org/abs/2104.10442
Autor:
Bochao Wang, Wayne Zhang, Liyang Liu, Zhanghui Kuang, Jing-Hao Xue, Qingmin Liao, Yimin Chen, Wenming Yang
Publikováno v:
IEEE transactions on neural networks and learning systems. 33(8)
Object detection has made enormous progress and has been widely used in many applications. However, it performs poorly when only limited training data is available for novel classes that the model has never seen before. Most existing approaches solve
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::6fe570bb2c43022b217f309484309b08
https://pubmed.ncbi.nlm.nih.gov/33523819
https://pubmed.ncbi.nlm.nih.gov/33523819
Autor:
Tsui Hin Lin, Jianyong Chen, Tong Gao, Xiaoyu Yue, Hongbin Sun, Kai Chen, Wayne Zhang, Huaqiang Wei, Wenwei Zhang, Zhanghui Kuang, Dahua Lin, Zhizhong Li, Yiqin Zhu
Publikováno v:
ACM Multimedia
We present MMOCR-an open-source toolbox which provides a comprehensive pipeline for text detection and recognition, as well as their downstream tasks such as named entity recognition and key information extraction. MMOCR implements 14 state-of-the-ar
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::1999db86661f580022685f5f02a91642
Most weakly supervised semantic segmentation (WSSS) methods follow the pipeline that generates pseudo-masks initially and trains the segmentation model with the pseudo-masks in fully supervised manner after. However, we find some matters related to t
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::54c6ed6c19b563bf043bd4491be91e90
Publikováno v:
IEEE transactions on neural networks and learning systems. 32(6)
Elastic weight consolidation (EWC) has been successfully applied for general incremental learning to overcome the catastrophic forgetting issue. It adaptively constrains each parameter of the new model not to deviate much from its counterpart in the
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::134f86234bc8212315fc2160d825f3fe
https://pubmed.ncbi.nlm.nih.gov/32598286
https://pubmed.ncbi.nlm.nih.gov/32598286
Publikováno v:
Computer Vision – ECCV 2020 ISBN: 9783030585945
ECCV (25)
ECCV (25)
Video action detection approaches usually conduct actor-centric action recognition over RoI-pooled features following the standard pipeline of Faster-RCNN. In this work, we first empirically find the recognition accuracy is highly correlated with the
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::4b637db080168e47be019cdc70cfbc35
https://doi.org/10.1007/978-3-030-58595-2_27
https://doi.org/10.1007/978-3-030-58595-2_27
Publikováno v:
Computer Vision – ECCV 2020 ISBN: 9783030585280
ECCV (19)
ECCV (19)
The attention-based encoder-decoder framework has recently achieved impressive results for scene text recognition, and many variants have emerged with improvements in recognition quality. However, it performs poorly on contextless texts (e.g., random
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::586608b70d274d9ef7f7d6cdae1fdca2
https://doi.org/10.1007/978-3-030-58529-7_9
https://doi.org/10.1007/978-3-030-58529-7_9