Search

Your search keyword '"Liu, Wenxian"' showing total 555 results
Start Over Author "Liu, Wenxian"
555 results on '"Liu, Wenxian"'

5. AutoShot: A Short Video Dataset and State-of-the-Art Shot Boundary Detection

Author

Zhu, Wentao, Huang, Yufang, Xie, Xiufeng, Liu, Wenxian, Deng, Jincan, Zhang, Debing, Wang, Zhangyang, and Liu, Ji

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence, Computer Science - Machine Learning, Computer Science - Multimedia, Computer Science - Neural and Evolutionary Computing
Abstract

The short-form videos have explosive popularity and have dominated the new social media trends. Prevailing short-video platforms,~\textit{e.g.}, Kuaishou (Kwai), TikTok, Instagram Reels, and YouTube Shorts, have changed the way we consume and create content. For video content creation and understanding, the shot boundary detection (SBD) is one of the most essential components in various scenarios. In this work, we release a new public Short video sHot bOundary deTection dataset, named SHOT, consisting of 853 complete short videos and 11,606 shot annotations, with 2,716 high quality shot boundary annotations in 200 test videos. Leveraging this new data wealth, we propose to optimize the model design for video SBD, by conducting neural architecture search in a search space encapsulating various advanced 3D ConvNets and Transformers. Our proposed approach, named AutoShot, achieves higher F1 scores than previous state-of-the-art approaches, e.g., outperforming TransNetV2 by 4.2%, when being derived and evaluated on our newly constructed SHOT dataset. Moreover, to validate the generalizability of the AutoShot architecture, we directly evaluate it on another three public datasets: ClipShots, BBC and RAI, and the F1 scores of AutoShot outperform previous state-of-the-art approaches by 1.1%, 0.9% and 1.2%, respectively. The SHOT dataset and code can be found in https://github.com/wentaozhu/AutoShot.git ., Comment: 10 pages, 5 figures, 3 tables, in CVPR 2023; Top-1 solution for scene / shot boundary detection https://paperswithcode.com/paper/autoshot-a-short-video-dataset-and-state-of

Published
2023

26. S‐Block Metal Mg‐Mediated Co─N─C as Efficient Oxygen Electrocatalyst for Durable and Temperature‐Adapted Zn–Air Batteries.

Author

Wang, Henan, Niu, Xinxin, Liu, Wenxian, Yin, Ruilian, Dai, Jiale, Guo, Wei, Kong, Chao, Ma, Lu, Ding, Xia, Wu, Fangfang, Shi, Wenhui, Deng, Tianqi, and Cao, Xiehong

Subjects
ORBITAL hybridization, OXYGEN reduction, DENSITY functional theory, TRANSITION metals, ACTIVATION energy
Abstract

In the quest to enhance Zn–air batteries (ZABs) for operating across a wide spectrum of temperatures, synthesizing robust oxygen electrocatalysts is paramount. Conventional strategies focusing on orbital hybridization of d–d and p‐d aim to moderate the excessive interaction between the d‐band of the transition metal active site and oxygen intermediate, yet often yield suboptimal performance. Herein, an innovative s‐block metal modulation is reported to refine the electronic structure and catalytic behavior of Co─NC catalysts. Employing density functional theory (DFT) calculations, it is revealed that incorporating Mg markedly depresses the d‐band center of Co sites, thereby fine‐tuning the adsorption energy of the oxygen reduction reaction (ORR) intermediate. Consequently, the Mg‐modified Co─NC catalyst (MgCo─NC) unveils remarkable intrinsic ORR activity with a significantly reduced activation energy (Ea) of 10.0 kJ mol−1, outstripping the performance of both Co─NC (17.6 kJ mol−1), benchmark Pt/C (15.9 kJ mol−1), and many recent reports. Moreover, ZABs outfitted with the finely tuned Mg0.1Co0.9─NC realize a formidable power density of 157.0 mW cm−2, paired with an extremely long cycle life of 1700 h, and an exceptionally minimal voltage gap decay rate of 0.006 mV h−1. Further, the Mg0.1Co0.9─NC‐based flexible ZAB presents a mere 2% specific capacity degradation when the temperature fluctuates from 25 to −20 °C, underscoring its robustness and suitability for practical deployment in diverse environmental conditions. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

35. Geometry-aware Similarity Learning on SPD Manifolds for Visual Recognition

Author

Huang, Zhiwu, Wang, Ruiping, Li, Xianqiu, Liu, Wenxian, Shan, Shiguang, Van Gool, Luc, and Chen, Xilin

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Symmetric Positive Definite (SPD) matrices have been widely used for data representation in many visual recognition tasks. The success mainly attributes to learning discriminative SPD matrices with encoding the Riemannian geometry of the underlying SPD manifold. In this paper, we propose a geometry-aware SPD similarity learning (SPDSL) framework to learn discriminative SPD features by directly pursuing manifold-manifold transformation matrix of column full-rank. Specifically, by exploiting the Riemannian geometry of the manifold of fixed-rank Positive Semidefinite (PSD) matrices, we present a new solution to reduce optimizing over the space of column full-rank transformation matrices to optimizing on the PSD manifold which has a well-established Riemannian structure. Under this solution, we exploit a new supervised SPD similarity learning technique to learn the transformation by regressing the similarities of selected SPD data pairs to their ground-truth similarities on the target SPD manifold. To optimize the proposed objective function, we further derive an algorithm on the PSD manifold. Evaluations on three visual classification tasks show the advantages of the proposed approach over the existing SPD-based discriminant learning methods.

Published
2016

40. Chromosome‐level genome assembly of Pedicularis kansuensis illuminates genome evolution of facultative parasitic plant.

Author

Fang, Longfa, Li, Mingyu, Zhang, Jia, Jia, Chenglin, Qiang, Yuqing, He, Xiaojuan, Liu, Tao, Zhou, Qiang, Luo, Dong, Han, Yuling, Li, Zhen, Liu, Wenxian, Yang, Yongzhi, Liu, Jianquan, and Liu, Zhipeng

Subjects
HORIZONTAL gene transfer, PARASITIC plants, GENOME size, GENOMES, PLANT genomes
Abstract

Parasitic plants have a heterotrophic lifestyle, in which they withdraw all or part of their nutrients from their host through the haustorium. Despite the release of many draft genomes of parasitic plants, the genome evolution related to the parasitism feature of facultative parasites remains largely unknown. In this study, we present a high‐quality chromosomal‐level genome assembly for the facultative parasite Pedicularis kansuensis (Orobanchaceae), which invades both legume and grass host species in degraded grasslands on the Qinghai‐Tibet Plateau. This species has the largest genome size compared with other parasitic species, and expansions of long terminal repeat retrotransposons accounting for 62.37% of the assembly greatly contributed to the genome size expansion of this species. A total of 42,782 genes were annotated, and the patterns of gene loss in P. kansuensis differed from other parasitic species. We also found many mobile mRNAs between P. kansuensis and one of its host species, but these mobile mRNAs could not compensate for the functional losses of missing genes in P. kansuensis. In addition, we identified nine horizontal gene transfer (HGT) events from rosids and monocots, as well as one single‐gene duplication events from HGT genes, which differ distinctly from that of other parasitic species. Furthermore, we found evidence for HGT through transferring genomic fragments from phylogenetically remote host species. Taken together, these findings provide genomic insights into the evolution of facultative parasites and broaden our understanding of the diversified genome evolution in parasitic plants and the molecular mechanisms of plant parasitism. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

41. Harnessing the Potential of Zn Powder Anodes: Innovations and Future Directions in Aqueous Zinc‐Ion Batteries.

Author

Liu, Dongshu, Xu, Xilian, Meng, Shibo, Li, Jun, Feng, Jinxiu, Chen, Yuchao, Liu, Wenxian, Wu, Fangfang, Shi, Wenhui, and Cao, Xiehong

Subjects
ANODES, POWDERS, THREE-dimensional printing, DENDRITES, STORAGE batteries, ZINC powder, AQUEOUS electrolytes
Abstract

The pursuit of alternative anode materials featuring inexpensive, efficient, and adaptable is shaping the future of rechargeable zinc‐ion batteries. The Zn powder anodes stand out among contenders due to their cost‐effectiveness, exceptional processability, and adjustability. However, their widespread use is hampered by significant challenges, including volume expansion and dendrite growth. To counter these issues, a gamut of optimization strategies for Zn powder anodes has been explored and developed. This review systematically encapsulates these research findings, offering a comprehensive understanding of various enhancement methods, such as three‐dimensional printing, in‐situ surface engineering, scalable electrostatic self‐assembly, and epoxy oligomer binding. These methodologies are discussed in detail in the context of their unique synthesis methods and underlying mechanisms. The paper concludes by outlining prospective research trajectories aimed at further optimizing the use of Zn powder anodes in aqueous zinc‐ion batteries, thereby illuminating potential avenues for future exploration and development. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results