Zhengwei Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhengwei Shen has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Changmiao Wang
Shenzhen Research Institute of Big Data, Shenzhen, China
3 shared publications - Feiwei Qin
Hangzhou Dianzi University
3 shared publications - Ruiquan Ge
Hangzhou Dianzi University
3 shared publications Kai Zhang
2 shared publications- Ahmed Elazab
Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China
1 shared publication Fei Lin
1 shared publication
Latest Publications
- Top Cited Papers
- 10 January 2025
- Published by Elsevier in Knowledge-Based Systems
- Vol. 310
InfraFFN: A Feature Fusion Network leveraging dual-path convolution and self-attention for infrared image super-resolution
- 10 January 2025
- Published by Elsevier in Knowledge-Based Systems
- Vol. 310
Top Cited PapersInfraFFN: A Feature Fusion Network leveraging dual-path convolution and self-attention for infrared image super-resolution
- Open Access
- 10 October 2024
- Published by Elsevier in Alexandria Engineering Journal
- Vol. 110, 310-321
IDTransformer: Infrared image denoising method based on convolutional transposed self-attention
- 10 October 2024
- Published by Elsevier in Alexandria Engineering Journal
- Vol. 110, 310-321
Open AccessIDTransformer: Infrared image denoising method based on convolutional transposed self-attention
- Open Access
- 18 March 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (6) , 899
Boundary-Match U-Shaped Temporal Convolutional Network for Vulgar Action Segmentation
The advent of deep learning has provided solutions to many challenges posed by the Internet. However, efficient localization and recognition of vulgar segments within videos remain formidable tasks. This difficulty arises from the blurring of spatial features in vulgar actions, which can render them indistinguishable from general actions. Furthermore, issues of boundary ambiguity and over-segmentation complicate the segmentation of vulgar actions. To address these issues, we present the Boundary-Match U-shaped Temporal Convolutional Network (BMUTCN), a novel approach for the segmentation of vulgar actions. The BMUTCN employs a U-shaped architecture within an encoder–decoder temporal convolutional network to bolster feature recognition by leveraging the context of the video. Additionally, we introduce a boundary-match map that fuses action boundary inform ation with greater precision for frames that exhibit ambiguous boundaries. Moreover, we propose an adaptive internal block suppression technique, which substantially mitigates over-segmentation errors while preserving accuracy. Our methodology, tested across several public datasets as well as a bespoke vulgar dataset, has demonstrated state-of-the-art performance on the latter.
- 18 March 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (6) , 899
Open AccessBoundary-Match U-Shaped Temporal Convolutional Network for Vulgar Action Segmentation
The advent of deep learning has provided solutions to many challenges posed by the Internet. However, efficient localization and recognition of vulgar segments within videos remain formidable tasks. This difficulty arises from the blurring of spatial features in vulgar actions, which can render them indistinguishable from general actions. Furthermore, issues of boundary ambiguity and over-segmentation complicate the segmentation of vulgar actions. To address these issues, we present the Boundary-Match U-shaped Temporal Convolutional Network (BMUTCN), a novel approach for the segmentation of vulgar actions. The BMUTCN employs a U-shaped architecture within an encoder–decoder temporal convolutional network to bolster feature recognition by leveraging the context of the video. Additionally, we introduce a boundary-match map that fuses action boundary inform ation with greater precision for frames that exhibit ambiguous boundaries. Moreover, we propose an adaptive internal block suppression technique, which substantially mitigates over-segmentation errors while preserving accuracy. Our methodology, tested across several public datasets as well as a bespoke vulgar dataset, has demonstrated state-of-the-art performance on the latter.