Zobrazeno 1 - 10
of 14
pro vyhledávání: '"Matt Feiszli"'
Publikováno v:
AAAI
University of Technology Sydney
University of Technology Sydney
Typical video classification methods often divide a video into short clips, do inference on each clip independently, then aggregate the clip-level predictions to generate the video-level results. However, processing visually similar clips independent
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783031198328
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::62812e1488e820b6a50b4b4c6aa0f240
https://doi.org/10.1007/978-3-031-19833-5_41
https://doi.org/10.1007/978-3-031-19833-5_41
Autor:
Jitendra Malik, Haoqi Fan, Ross Girshick, Aaron Adcock, Matt Feiszli, Haichuan Yang, Bo Xiong, Nikhila Ravi, Yanghao Li, Christoph Feichtenhofer, Tullie Murrell, Heng Wang, Kalyan Vasudev Alwala, Meng Li, Wan-Yen Lo, Yilei Li
Publikováno v:
ACM Multimedia
We introduce PyTorchVideo, an open-source deep-learning library that provides a rich set of modular, efficient, and reproducible components for a variety of video understanding tasks, including classification, detection, self-supervised learning, and
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::41950dc1e1f1cc0f6f48cfaea938e168
http://arxiv.org/abs/2111.09887
http://arxiv.org/abs/2111.09887
Autor:
Laura Sevilla-Lara, Vedanuj Goswami, Lorenzo Torresani, Zhicheng Yan, Shengxin Zha, Matt Feiszli
Publikováno v:
WACV
Understanding temporal information and how the visual world changes over time, is a fundamental ability of intelligent systems. In video understanding, temporal information is at the core of many current challenges, including compression, efficient i
Autor:
Krishna Kumar Singh, Matt Feiszli, Kristen Grauman, Yong Jae Lee, Dhruv Mahajan, Deepti Ghadiyaram
Publikováno v:
CVPR
Existing models often leverage co-occurrences between objects and their context to improve recognition accuracy. However, strongly relying on context risks a model's generalizability, especially when typical co-occurrence patterns are absent. This wo
Publikováno v:
Computer Vision – ECCV 2020 ISBN: 9783030585471
ECCV (4)
University of Technology Sydney
ECCV (4)
University of Technology Sydney
In this paper, we study an intermediate form of supervision, i.e., single-frame supervision, for temporal action localization (TAL). To obtain the single-frame supervision, the annotators are asked to identify only a single frame within the temporal
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::5de4c1848d2cd0d3a54d98df87c3ae89
https://doi.org/10.1007/978-3-030-58548-8_25
https://doi.org/10.1007/978-3-030-58548-8_25
Autor:
Akil Narayan, Matt Feiszli
Publikováno v:
SIAM Journal on Imaging Sciences. 10:1322-1345
We propose an optimization algorithm for computing geodesics on the universal Teichmuller space $T(1)$ in the Weil--Petersson (WP) metric. Another realization for T(1) is the space of planar shapes, modulo translation and scale, and thus our algorith
Publikováno v:
CVPR
Motion is a salient cue to recognize actions in video. Modern action recognition models leverage motion information either explicitly by using optical flow as input or implicitly by means of 3D convolutional filters that simultaneously capture appear
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::97d0eebfc39d5cbd252fd0f4a9beb2d7
http://arxiv.org/abs/1906.03349
http://arxiv.org/abs/1906.03349
Publikováno v:
CVPR
Consider end-to-end training of a multi-modal vs. a single-modal network on a task with multiple input modalities: the multi-modal network receives more information, so it should match or outperform its single-modal counterpart. In our experiments, h
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7d4b62e2db6b5a646b036045f2649cc9
Publikováno v:
ICCV
Group convolution has been shown to offer great computational savings in various 2D convolutional architectures for image classification. It is natural to ask: 1) if group convolution can help to alleviate the high computational cost of video classif
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::4189e75666b6f5def0822a6b54b315db