Search

Your search keyword '"Liao, Yujun"' showing total 156 results
Start Over Author "Liao, Yujun"
156 results on '"Liao, Yujun"'

3. Self-Supervised Learning Across Domains

Author

Bucci, Silvia, D'Innocente, Antonio, Liao, Yujun, Carlucci, Fabio Maria, Caputo, Barbara, and Tommasi, Tatiana

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Human adaptability relies crucially on learning and merging knowledge from both supervised and unsupervised tasks: the parents point out few important concepts, but then the children fill in the gaps on their own. This is particularly effective, because supervised learning can never be exhaustive and thus learning autonomously allows to discover invariances and regularities that help to generalize. In this paper we propose to apply a similar approach to the problem of object recognition across domains: our model learns the semantic labels in a supervised fashion, and broadens its understanding of the data by learning from self-supervised signals on the same images. This secondary task helps the network to focus on object shapes, learning concepts like spatial orientation and part correlation, while acting as a regularizer for the classification task over multiple visual domains. Extensive experiments confirm our intuition and show that our multi-task method combining supervised and self-supervised knowledge shows competitive results with respect to more complex domain generalization and adaptation solutions. It also proves its potential in the novel and challenging predictive and partial domain adaptation scenarios., Comment: Accepted at IEEE T-PAMI

Published
2020

36. Self-Supervised Learning Across Domains.

Author

Bucci, Silvia, D'Innocente, Antonio, Liao, Yujun, Carlucci, Fabio M., Caputo, Barbara, and Tommasi, Tatiana

Subjects
SUPERVISED learning, OBJECT recognition (Computer vision), SPATIAL orientation, CONCEPT learning, LEARNING, IMAGE recognition (Computer vision)
Abstract

Human adaptability relies crucially on learning and merging knowledge from both supervised and unsupervised tasks: the parents point out few important concepts, but then the children fill in the gaps on their own. This is particularly effective, because supervised learning can never be exhaustive and thus learning autonomously allows to discover invariances and regularities that help to generalize. In this paper we propose to apply a similar approach to the problem of object recognition across domains: our model learns the semantic labels in a supervised fashion, and broadens its understanding of the data by learning from self-supervised signals on the same images. This secondary task helps the network to focus on object shapes, learning concepts like spatial orientation and part correlation, while acting as a regularizer for the classification task over multiple visual domains. Extensive experiments confirm our intuition and show that our multi-task method, combining supervised and self-supervised knowledge, provides competitive results with respect to more complex domain generalization and adaptation solutions. It also proves its potential in the novel and challenging predictive and partial domain adaptation scenarios. [ABSTRACT FROM AUTHOR]

Published
2022
Full Text
View/download PDF

37. A Study of Blended Learning Using the Smart Class Teaching Module on Psychosocial Dysfunction Course During the Training of Undergraduate Occupational Therapy Students in China

Author

Lin, Qiang, primary, Qiu, Yaxian, additional, Zhang, Yanni, additional, Zheng, Yuxing, additional, Zhang, Lingzi, additional, Liang, Junjie, additional, Anniwaer, Yilifate, additional, Liao, Yujun, additional, Jiang, Yueyi, additional, Lang, Shijuan, additional, Zhuang, Wanqi, additional, Luo, Siyi, additional, Zheng, Meifeng, additional, Yan, QingHua, additional, Chen, Ke, additional, Chen, Qiuxia, additional, and Ou, Haining, additional

Published
2021
Full Text
View/download PDF

50. Preoperative evaluation of moyamoya spontaneous anastomosis of combined revascularization donor vessels in adults by duplex ultrasonography.

Author

Chen, Li, Xu, Bin, Wang, Yong, Liao, Yujun, Pan, Huiwen, and Wang, Yi

Subjects
ULTRASONIC imaging, ANGIOGRAPHY, ADULTS, AGE groups, MEDICAL radiography
Abstract

Objective: To evaluate the preoperative diagnostic value of duplex ultrasonography in moyamoya spontaneous anastomosis of combined revascularization donor vessels in adults. Methods: A total of 99 preoperative adult patients who underwent superficial temporal artery-to-middle cerebral artery (STA-MCA) anastomosis were retrospectively analyzed. Each side of the cerebral hemisphere was examined as a separate procedure. A total of 198 cerebral hemispheres were divided into three groups: a collateral, non-collateral, and control group based on digital subtraction angiography (DSA). Hemodynamic parameters, including peak systolic velocity (PSV), end diastolic velocity (EDV), and resistance index (RI) were analyzed. Results: There were only four of cases (5%, 4/198) of STA spontaneous anastomosis, whereas those of maxillary artery (MA) anastomosis were 44 (23.7%, 44/186). Compared with the control group, MA PSV and EDV of the collateral group increased significantly, while RI decreased significantly (p < .05). The area under the curve (AUC) of MA RI was 0.654. As a predictor of MA spontaneous anastomosis, duplex ultrasonography had high specificity but poor sensitivity. In collateral group, PSV and EDV detected two weeks post-surgery were significantly higher than those detected pre-operatively (PSV: p = .018, EDV: p = .025). By contrast, there were no significant difference of the PSV and EDV detected six months post-surgery compared with pre-operation (PSV: p = .450, EDV: p = .099). Additionally, MA RI in two weeks after the surgery was comparable with preoperative values. Conclusions: Duplex ultrasonography could be applied to evaluate the adult moyamoya spontaneous anastomoses of MA preoperatively. Despite its poor sensitivity, this diagnostic modality is still reliable and specific. STA-MCA anastomosis combined with EDMS did not affect MA pre-operative spontaneous anastomosis. [ABSTRACT FROM AUTHOR]

Published
2018
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results