Filters








29,781 Hits in 5.9 sec

Bottom-Up Temporal Action Localization with Mutual Regularization [article]

Peisen Zhao, Lingxi Xie, Chen Ju, Ya Zhang, Yanfeng Wang, Qi Tian
2021 arXiv   pre-print
The proposed regularization also generalizes to other TAL methods (e.g., TSA-Net and PGCN). code: https://github.com/PeisenZhao/Bottom-Up-TAL-with-MR  ...  Recently, temporal action localization (TAL), i.e., finding specific action segments in untrimmed videos, has attracted increasing attentions of the computer vision community.  ...  Moreover, we show that the proposed mutual regularization is independent of the temporal action localization framework.  ... 
arXiv:2002.07358v3 fatcat:nhludhkxxvgbrest2u6ekpywem

Feature binding as neuron synchronization: quantum aspects

F. Tito Arecchi
2005 Brazilian journal of physics  
The winning pattern must be confirmed by some matching procedure between bottom-up and top-down signals.  ...  action.  ...  One (bottom-up) comes from the sensory detectors via the early stages which classify elementary features.  ... 
doi:10.1590/s0103-97332005000200008 fatcat:epjfjv6lerhwvixkbh2pk5lfwm

Visual pattern discovery in image and video data: a brief survey

Hongxing Wang, Gangqiang Zhao, Junsong Yuan
2013 Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery  
The bottom-up pattern discovery method starts with unordered visual primitives followed by merging the primitives until larger visual patterns are found.  ...  We categorize the existing methods into two groups: bottom-up pattern discovery and top-down pattern modeling.  ...  Liu and Chen 31 extend topic models from still images to videos with a temporal model integrated.  ... 
doi:10.1002/widm.1110 fatcat:skjnmv5njfdtxc3erl4r2txqri

Intra- and Inter-Action Understanding via Temporal Action Parsing [article]

Dian Shao, Yue Zhao, Bo Dai, Dahua Lin
2020 arXiv   pre-print
Towards this goal, we construct TAPOS, a new dataset developed on sport videos with manual annotations of sub-actions, and conduct a study on temporal action parsing on top.  ...  Our study shows that a sport activity usually consists of multiple sub-actions and that the awareness of such temporal structures is beneficial to action recognition.  ...  We use two losses, i.e. a local loss to promote agreement between frames within a sub-action while suppressing that across sub-actions and a global loss to predict action label as a regularization.  ... 
arXiv:2005.10229v1 fatcat:n4zcmoo7mfc2xhje3pm32tb5ka

Locating Order-Disorder Phase Transition in a Cardiac System [article]

Hiroshi Ashikaga, Ameneh Asgari-Targhi
2017 arXiv   pre-print
Communication analysis between the driving component and each component of the system reveals that channel capacity, mutual information and transfer entropy can locate the wavebreaks.  ...  Cardiac action potential. Three cardiac action potentials in response to regular stimuli with an external current ('Stim') are shown.  ...  At BCL=300 msec, one stimulus generates one action potential, and APD is uniform spatially and temporally (APD=185 msec, Figure 3A ).  ... 
arXiv:1708.03990v2 fatcat:ft4kxtdfzjcyfaqoalbhwi33om

COMPLEXITY OF PERCEPTUAL PROCESSES

F. TITO ARECCHI
2005 Complexity, Metastability and Nonextensivity  
An external stimulus spreads over an assembly of coupled neurons, building up a corresponding collective state.  ...  It is conjectured that they are related to the details of the perceptual chain rather than depending on Planck's action. This is where the abstract should be placed.  ...  The winning pattern must be confirmed by some matching procedure between bottom-up and top-down signals.  ... 
doi:10.1142/9789812701558_0030 fatcat:ivm6j5ym6feazcd4ioeb6t3kxu

Activity recognition by learning structural and pairwise mid-level features using random forest

Jie Hu, Yu Kong, Yun Fu
2013 2013 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG)  
Experimental results show that our method achieves promising performance on three available action and facial expression datasets.  ...  Our model consists of two separate parts, spatial part and temporal part, which are employed to capture the distinctive characteristics in spatial and temporal domains of activity analysis.  ...  Local features in spatial neighborhood make up the specific appearance of the action for each temporal interval.  ... 
doi:10.1109/fg.2013.6553706 dblp:conf/fgr/HuKF13 fatcat:fifvvznno5hwxmvaa4r2gakocu

Study of convective heat transfer in flat plate solar collectors

Yedilkhan Amirgaliyev, Murat Kunelbayev, Aliya Kalizhanova, Beibut Amirgaliyev, Ainur Kozbakova, Omirlan Auelbekov, Nazbek Kataev, N. Bardis
2019 ITM Web of Conferences  
In each case the lines are with a tip 1/3, as in the research herein only the laminar conditions have been considered.  ...  plate solar collectors, as it is seen from the analysis on research of the heat transfer by a circular and flat tubes upon conforming the forced and free convections, placed vertically or horizontally with  ...  action of the forced and free convection can be broken down into three groups in compliance with the three main types of flow: а) flow in vertical tubes bottom-up at heating and top-bottom at cooling;  ... 
doi:10.1051/itmconf/20192401009 fatcat:i5nhll76o5d53jpzfbe3zi25wm

Spatio-Temporal Co-Occurrence Characterizations For Human Action Classification

Aznul Qalid Md Sabri, Jacques Boonaert, Erma Rahayu Mohd Faizal Abdullah, Ali Mohammed Mansoor
2017 Malaysian Journal of Computer Science  
Many state-ofthe-arts approaches involve the usage of bag-of-video-words with spatio-temporal local features to construct characterizations for human actions.  ...  Next, a spatio-temporal co-occurrence technique is applied to extract co-occurrence information between labeled local features. Novel characterizations for human actions are then constructed.  ...  During the video-words selection process, a greedy algorithm based on a bottom-up pair wise merging procedure is performed. The algorithm starts by treating each element of  as a singleton cluster.  ... 
doi:10.22452/mjcs.vol30no3.1 fatcat:hui3mh3bardorg4ucyt7m4u7ue

Saving the mutual manipulability account of constitutive relevance

Beate Krickel
2018 Studies in History and Philosophy of Science Part A  
My approach is based on a causal interpretation of mutual manipulability, where manipulability is interpreted as a causal relation between the mechanism's components and temporal parts of the phenomenon  ...  The most prominent approach of how to understand this constitution relation is Carl Craver's mutual manipulability approach to constitutive relevance.  ...  This would correspond to an ideal* bottom-up intervention.  ... 
doi:10.1016/j.shpsa.2018.01.003 pmid:29609776 fatcat:tkjwhddzq5arfpak7vwirhqcaq

Temporal Localization of Actions with Actoms

A. Gaidon, Z. Harchaoui, C. Schmid
2013 IEEE Transactions on Pattern Analysis and Machine Intelligence  
We show that our ASM method outperforms the current state of the art in temporal action localization, as well as baselines that localize actions with a sliding window method.  ...  We address the problem of localizing actions, such as opening a door, in hours of challenging video data.  ...  Fig. 7 : 7 Frames of the top 5 actions localized with ASM for "Drinking" (top row) and "Open Door" (bottom row).  ... 
doi:10.1109/tpami.2013.65 pmid:24051735 fatcat:s7c6dth3qrarje3wqhkaqyehem

Spatio-temporal Co-Occurrence Characterizations for Human Action Classification [article]

Aznul Qalid Md Sabri, Jacques Boonaert, Erma Rahayu Mohd Faizal Abdullah, Ali Mohammed Mansoor
2016 arXiv   pre-print
Many state-of-the-arts approaches involve the usage of bag-of-video-words with spatio-temporal local features to construct characterizations for human actions.  ...  Next, a spatio-temporal co-occurrence technique is applied to extract co-occurrence information between labeled local features. Novel characterizations for human actions are then constructed.  ...  During the video-words selection process, a greedy algorithm based on a bottom-up pair wise merging procedure is performed. The algorithm starts by treating each element of  as a singleton cluster.  ... 
arXiv:1610.05174v1 fatcat:pq7qfnwihfgvlpsobrerollr3i

DCAN: Improving Temporal Action Detection via Dual Context Aggregation [article]

Guo Chen, Yin-Dong Zheng, Limin Wang, Tong Lu
2021 arXiv   pre-print
Temporal action detection aims to locate the boundaries of action in the video.  ...  proposals, thereby improving the performance of temporal action detection.  ...  Ablation Studies Ablation comparison with other "bottom-up" methods.  ... 
arXiv:2112.03612v1 fatcat:n73qmjmnonfhtlulu33te6kuea

Capturing Global Semantic Relationships for Facial Action Unit Recognition

Ziheng Wang, Yongqiang Li, Shangfei Wang, Qiang Ji
2013 2013 IEEE International Conference on Computer Vision  
Towards this goal, we build a hierarchical model that combines the bottom-level image features and the top-level AU relationships to jointly recognize AUs in a principled manner.  ...  In this paper we tackle the problem of facial action unit (AU) recognition by exploiting the complex semantic relationships among AUs, which carry crucial top-down information yet have not been thoroughly  ...  Bottom up -AU recognition from image features. The bottom two layers integrate the image shape or appearance features for AU recognition in the bottom-up direction.  ... 
doi:10.1109/iccv.2013.410 dblp:conf/iccv/WangLWJ13 fatcat:zgzol76gy5balc5onhqr74ubbi

Graph Constrained Data Representation Learning for Human Motion Segmentation [article]

Mariella Dimiccoli, Lluís Garrido, Guillem Rodriguez-Corominas, Herwig Wendt
2021 arXiv   pre-print
Second, we introduce a regularization term for this auxiliary data matrix that preserves the local geometrical structure present in the high-dimensional space.  ...  Our model is reminiscent of temporal subspace clustering, but presents two critical differences.  ...  Figure 4 : 4 Frobenius norm of the temporal second derivative (left) and difference with data X (right) of auxiliary repre-sentationX for different values of the regularization term L G and temporal regularization  ... 
arXiv:2107.13362v2 fatcat:ypfc43zvcjb55mce2rvgrlgfua
« Previous Showing results 1 — 15 out of 29,781 results