Chun-Xia Huang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Chun-Xia Huang has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Dongdong Zhao
College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou 310023, China
1 shared publication - Fengying Chen
Department of Nursing, Traditional Chinese Medicine Hospital of Kunshan, Kunshan 215300, Jiangsu Province, China
1 shared publication - Juan Wei
College of Electrical and Information Engineering, Hunan University, Changsha 410082, China
1 shared publication Lan Cai
Department of Nursing, Traditional Chinese Medicine Hospital of Kunshan, Kunshan 215300, Jiangsu Province, China
1 shared publicationLi-Qin Cai
The Fourth Ward of Surgery, Traditional Chinese Medicine Hospital of Kunshan, Kunshan 215300, Jiangsu Province, China
1 shared publication- Nan Gao
Zhejiang University of Technology
1 shared publication
Latest Publications
- Open Access
- 8 March 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 20 (6) , 1-21
Discriminative Action Snippet Propagation Network for Weakly Supervised Temporal Action Localization
Weakly supervised temporal action localization (WTAL) aims to classify and localize actions in untrimmed videos with only video-level labels. Recent studies have attempted to obtain more accurate temporal boundaries by exploiting latent action instances in ambiguous snippets or propagating representative action features. However, empirically handcrafted ambiguous snippet extraction and the imprecise alignment of representative snippet propagation lead to challenges in modeling the completeness of actions for these methods. In this article, we propose a Discriminative Action Snippet Propagation Network (DASP-Net) to accurately discover ambiguous snippets in videos and propagate discriminative instance-level features throughout the video for improving action completeness. Specifically, we introduce a novel discriminative feature propagation module for capturing the global contextual attention and propagating the action concept across the whole video by perceiving the discriminative action snippets with instance information from the same video. Simultaneously, we incorporate denoised pseudo-labels as supervision, where we correct the controversial prediction based on the feature space distribution during training, thereby alleviating false detection caused by noise background features. Furthermore, we design an ambiguous feature mining module, which maximizes the feature affinity information of action and background in ambiguous snippets to generate more accurate latent action and background snippets and learns more precise action instance boundaries through contrastive learning of action and background snippets. Extensive experiments show that DASP-Net achieves state-of-the-art results on THUMOS14 and ActivityNet1.2 datasets.
- 8 March 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 20 (6) , 1-21
Open AccessDiscriminative Action Snippet Propagation Network for Weakly Supervised Temporal Action Localization
Weakly supervised temporal action localization (WTAL) aims to classify and localize actions in untrimmed videos with only video-level labels. Recent studies have attempted to obtain more accurate temporal boundaries by exploiting latent action instances in ambiguous snippets or propagating representative action features. However, empirically handcrafted ambiguous snippet extraction and the imprecise alignment of representative snippet propagation lead to challenges in modeling the completeness of actions for these methods. In this article, we propose a Discriminative Action Snippet Propagation Network (DASP-Net) to accurately discover ambiguous snippets in videos and propagate discriminative instance-level features throughout the video for improving action completeness. Specifically, we introduce a novel discriminative feature propagation module for capturing the global contextual attention and propagating the action concept across the whole video by perceiving the discriminative action snippets with instance information from the same video. Simultaneously, we incorporate denoised pseudo-labels as supervision, where we correct the controversial prediction based on the feature space distribution during training, thereby alleviating false detection caused by noise background features. Furthermore, we design an ambiguous feature mining module, which maximizes the feature affinity information of action and background in ambiguous snippets to generate more accurate latent action and background snippets and learns more precise action instance boundaries through contrastive learning of action and background snippets. Extensive experiments show that DASP-Net achieves state-of-the-art results on THUMOS14 and ActivityNet1.2 datasets.
- Open Access
- 6 April 2022
- Published by Baishideng Publishing Group Inc. in World Journal of Clinical Cases
- Vol. 10 (10) , 3035-3046
Early warning prevention and control strategies to reduce perioperative venous thromboembolism in patients with gastrointestinal cancer
Venous thromboembolism (VTE) is a major cause of unexpected and perioperative in-hospital deaths. It is characterized by high morbidity, high mortality, high misdiagnosis rate, and high missed diagnosis rates. VTE is a common postoperative complication in cancer patients. VTE is preventable, and early identification of risk factors leading to VTE and appropriate early preventive actions can reduce its occurrence and mortality. Presently, there is no uniform standard for the prevention and control of VTE in clinical practice, and hospitals in China lack mature and effective protocols for the assessment, prevention, and treatment of VTE. To explore whether an early warning program could influence the occurrence of deep vein thrombosis (DVT) postoperatively. This is a comparative retrospective cohort study, which enrolled patients who underwent laparotomic or laparoscopic gastrointestinal tumor resection for gastrointestinal cancer between January 2016 and December 2019. Patients were divided into a control group and an early warning group depending on whether or not the early warning program was implemented. A venous thromboembolism prevention and control team was established. The outcomes included the occurrence of DVT, the correct rate of VTE assessment, the coagulation indicators, and the mastery of VTE knowledge by the nurses. A total of 264 patients were included in this study, with 128 patients in the control group and 136 patients in the early warning group. The occurrence rate of DVT in the early warning group was 6.6% (9/136), compared with 14.1% (18/128) in the control group (P < 0.05). The correct rates of VTE risk assessment by the nurses and standard implementation rate of VTE preventive measures were 86.8% vs 65.6% and 80.2% vs 57.8% in early warning and control groups, respectively (all P < 0.001). The independent factors associated with postoperative DVT occurrence were age (OR = 1.083, 95%CI: 1.070-3.265, P = 0.032), Hyperlipidemia (OR = 1.127, 95%CI: 1.139-2.564, P = 0.042), preoperative high VTE risk (OR = 2.131, 95%CI: 1.085-5.178, P = 0.001), time of operation (OR = 2.268, 95%CI: 2.005-5.546, P = 0.026) and not adoption of early warning prevention (OR = 3.747, 95%CI: 1.523-6.956, P = 0.017). The early warning strategy was independently associated with the decreasing occurrence of VTE, and it might be suitable for protection from VTE in patients undergoing gastrointestinal cancer surgery.
- 6 April 2022
- Published by Baishideng Publishing Group Inc. in World Journal of Clinical Cases
- Vol. 10 (10) , 3035-3046
Open AccessEarly warning prevention and control strategies to reduce perioperative venous thromboembolism in patients with gastrointestinal cancer
Venous thromboembolism (VTE) is a major cause of unexpected and perioperative in-hospital deaths. It is characterized by high morbidity, high mortality, high misdiagnosis rate, and high missed diagnosis rates. VTE is a common postoperative complication in cancer patients. VTE is preventable, and early identification of risk factors leading to VTE and appropriate early preventive actions can reduce its occurrence and mortality. Presently, there is no uniform standard for the prevention and control of VTE in clinical practice, and hospitals in China lack mature and effective protocols for the assessment, prevention, and treatment of VTE. To explore whether an early warning program could influence the occurrence of deep vein thrombosis (DVT) postoperatively. This is a comparative retrospective cohort study, which enrolled patients who underwent laparotomic or laparoscopic gastrointestinal tumor resection for gastrointestinal cancer between January 2016 and December 2019. Patients were divided into a control group and an early warning group depending on whether or not the early warning program was implemented. A venous thromboembolism prevention and control team was established. The outcomes included the occurrence of DVT, the correct rate of VTE assessment, the coagulation indicators, and the mastery of VTE knowledge by the nurses. A total of 264 patients were included in this study, with 128 patients in the control group and 136 patients in the early warning group. The occurrence rate of DVT in the early warning group was 6.6% (9/136), compared with 14.1% (18/128) in the control group (P < 0.05). The correct rates of VTE risk assessment by the nurses and standard implementation rate of VTE preventive measures were 86.8% vs 65.6% and 80.2% vs 57.8% in early warning and control groups, respectively (all P < 0.001). The independent factors associated with postoperative DVT occurrence were age (OR = 1.083, 95%CI: 1.070-3.265, P = 0.032), Hyperlipidemia (OR = 1.127, 95%CI: 1.139-2.564, P = 0.042), preoperative high VTE risk (OR = 2.131, 95%CI: 1.085-5.178, P = 0.001), time of operation (OR = 2.268, 95%CI: 2.005-5.546, P = 0.026) and not adoption of early warning prevention (OR = 3.747, 95%CI: 1.523-6.956, P = 0.017). The early warning strategy was independently associated with the decreasing occurrence of VTE, and it might be suitable for protection from VTE in patients undergoing gastrointestinal cancer surgery.