Online recognition of unsegmented actions with hierarchical SOM architecture

被引:5
作者
Gharaee, Zahra [1 ,2 ]
机构
[1] Linkoping Univ, Comp Vis Lab CVL, Linkoping, Sweden
[2] Lund Univ, Dept Philosophy & Cognit Sci, Helgonavagen 3, S-22100 Lund, Sweden
关键词
Action recognition and segmentation; Self-organizing neural networks; Cognitive architecture; Online performance; Hierarchical models; DEEP NEURAL-NETWORK; POSE; FEATURES; LATENCY; MOTION; MODEL;
D O I
10.1007/s10339-020-00986-4
中图分类号
B84 [心理学];
学科分类号
04 ; 0402 ;
摘要
Automatic recognition of an online series of unsegmented actions requires a method for segmentation that determines when an action starts and when it ends. In this paper, a novel approach for recognizing unsegmented actions in online test experiments is proposed. The method uses self-organizing neural networks to build a three-layer cognitive architecture. The unique features of an action sequence are represented as a series of elicited key activations by the first-layer self-organizing map. An average length of a key activation vector is calculated for all action sequences in a training set and adjusted in learning trials to generate input patterns to the second-layer self-organizing map. The pattern vectors are clustered in the second layer, and the clusters are then labeled by an action identity in the third layer neural network. The experiment results show that although the performance drops slightly in online experiments compared to the offline tests, the ability of the proposed architecture to deal with the unsegmented action sequences as well as the online performance makes the system more plausible and practical in real-case scenarios.
引用
收藏
页码:77 / 91
页数:15
相关论文
共 52 条
[1]  
[Anonymous], 2011, CVPR 2011, DOI DOI 10.1109/CVPR.2011.5995316
[2]  
[Anonymous], 2007, Body, language and mind, Volume 1: Embodiment, DOI 10.1515/9783110207507.2.167
[3]  
[Anonymous], 2012, ROBUST 3D ACTION REC
[4]  
[Anonymous], 2016, PROC CVPR IEEE, DOI DOI 10.1109/CVPR.2016.214
[5]   Ikaros: Building cognitive models for robots [J].
Balkenius, Christian ;
Moren, Jan ;
Johansson, Birger ;
Johnsson, Magnus .
ADVANCED ENGINEERING INFORMATICS, 2010, 24 (01) :40-48
[6]   Predictive-Corrective Networks for Action Detection [J].
Dave, Achal ;
Russakovsky, Olga ;
Ramanan, Deva .
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :2067-2076
[7]   A Novel Manifold Regularized Online Semi-supervised Learning Model [J].
Ding, Shuguang ;
Xi, Xuanyang ;
Liu, Zhiyong ;
Qiao, Hong ;
Zhang, Bo .
COGNITIVE COMPUTATION, 2018, 10 (01) :49-61
[8]  
Dollar P., 2005, ICCCN, VVolume 2005, P65, DOI [DOI 10.1109/VSPETS.2005.1570899, 10.1109/VSPETS.2005.1570899]
[9]   Exploring the Trade-off Between Accuracy and Observational Latency in Action Recognition [J].
Ellis, Chris ;
Masood, Syed Zain ;
Tappen, Marshall F. ;
LaViola, Joseph J., Jr. ;
Sukthankar, Rahul .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 2013, 101 (03) :420-436
[10]   Using Conceptual Spaces to Model Actions and Events [J].
Gardenfors, Peter ;
Warglien, Massimo .
JOURNAL OF SEMANTICS, 2012, 29 (04) :487-519