scholarly journals Learning Spatio-Temporal Representation With Local and Global Diffusion

2019 ◽  
Author(s):  
Zhaofan Qiu ◽  
Ting Yao ◽  
Chong-Wah Ngo ◽  
Xinmei Tian ◽  
Tao Mei
Keyword(s):  
Temporal Representation ◽  
Global Diffusion ◽  
Spatio Temporal

Local Spatio-Temporal Representation Using the 3D Shearlet Transform

2018 ◽  
Vol 17 (1) ◽  
pp. 57-72
Author(s):  
Damiano Malafronte ◽  
Ernesto De Vito ◽  
Francesca Odone
Keyword(s):  
Shearlet Transform ◽  
Temporal Representation ◽  
Spatio Temporal

scholarly journals Data on spatio-temporal representation of mineral N fertilization and manure N application as well as ammonia volatilization in French regions for the crop year 2005/06

Data in Brief ◽  
2018 ◽  
Vol 21 ◽  
pp. 1119-1124
Author(s):  
Sophie Génermont ◽  
Maharavo Marie Julie Ramanantenasoa ◽  
Karine Dufosse ◽  
Olivier Maury ◽  
Catherine Mignolet ◽  
...  
Keyword(s):  
Ammonia Volatilization ◽  
N Fertilization ◽  
Mineral N ◽  
N Application ◽  
Temporal Representation ◽  
Spatio Temporal

scholarly journals Temporal Interlacing Network

2020 ◽  
Vol 34 (07) ◽  
pp. 11966-11973
Author(s):  
Hao Shao ◽  
Shengju Qian ◽  
Yu Liu
Keyword(s):  
State Of The Art ◽  
Temporal Information ◽  
Spatial Representations ◽  
The Past ◽  
Temporal Representation ◽  
Temporal Features ◽  
Temporal Models ◽  
Long Time ◽  
Temporal Model ◽  
Spatio Temporal

For a long time, the vision community tries to learn the spatio-temporal representation by combining convolutional neural network together with various temporal models, such as the families of Markov chain, optical flow, RNN and temporal convolution. However, these pipelines consume enormous computing resources due to the alternately learning process for spatial and temporal information. One natural question is whether we can embed the temporal information into the spatial one so the information in the two domains can be jointly learned once-only. In this work, we answer this question by presenting a simple yet powerful operator – temporal interlacing network (TIN). Instead of learning the temporal features, TIN fuses the two kinds of information by interlacing spatial representations from the past to the future, and vice versa. A differentiable interlacing target can be learned to control the interlacing process. In this way, a heavy temporal model is replaced by a simple interlacing operator. We theoretically prove that with a learnable interlacing target, TIN performs equivalently to the regularized temporal convolution network (r-TCN), but gains 4% more accuracy with 6x less latency on 6 challenging benchmarks. These results push the state-of-the-art performances of video understanding by a considerable margin. Not surprising, the ensemble model of the proposed TIN won the 1st place in the ICCV19 - Multi Moments in Time challenge. Code is made available to facilitate further research.1

Sign in / Sign up

Export Citation Format

Share Document