Real-Time Multi-scale Action Detection from 3D Skeleton Data. Sharaf, A., Torki, M., Hussein, M. E., & El-Saban, M. In 2015 IEEE Winter Conference on Applications of Computer Vision, pages 998–1005, 2015. doi abstract bibtex In this paper we introduce a real-time system for action detection. The system uses a small set of robust features extracted from 3D skeleton data. Features are effectively described based on the probability distribution of skeleton data. The descriptor computes a pyramid of sample covariance matrices and mean vectors to encode the relationship between the features. For handling the intra-class variations of actions, such as action temporal scale variations, the descriptor is computed using different window scales for each action. Discriminative elements of the descriptor are mined using feature selection. The system achieves accurate detection results on difficult unsegmented sequences. Experiments on MSRC-12 and G3D datasets show that the proposed system outperforms the state-of-the-art in detection accuracy with very low latency. To the best of our knowledge, we are the first to propose using multi-scale description in action detection from 3D skeleton data.
@inproceedings{sharaf_real-time_2015,
title = {Real-Time Multi-scale Action Detection from 3D Skeleton Data},
rights = {All rights reserved},
doi = {10.1109/WACV.2015.138},
abstract = {In this paper we introduce a real-time system for action detection. The system uses a small set of robust features extracted from 3D skeleton data. Features are effectively described based on the probability distribution of skeleton data. The descriptor computes a pyramid of sample covariance matrices and mean vectors to encode the relationship between the features. For handling the intra-class variations of actions, such as action temporal scale variations, the descriptor is computed using different window scales for each action. Discriminative elements of the descriptor are mined using feature selection. The system achieves accurate detection results on difficult unsegmented sequences. Experiments on {MSRC}-12 and G3D datasets show that the proposed system outperforms the state-of-the-art in detection accuracy with very low latency. To the best of our knowledge, we are the first to propose using multi-scale description in action detection from 3D skeleton data.},
eventtitle = {2015 {IEEE} Winter Conference on Applications of Computer Vision},
pages = {998--1005},
booktitle = {2015 {IEEE} Winter Conference on Applications of Computer Vision},
author = {Sharaf, A. and Torki, M. and Hussein, M. E. and El-Saban, M.},
year = {2015},
keywords = {3D skeleton data, covariance matrices, Detectors, feature extraction, Feature extraction, G3D datasets, Joints, mean vectors, {MSRC}-12, object detection, real-time multiscale action detection, real-time system, real-time systems, Real-time systems, Three-dimensional displays, vectors, Vectors},
file = {IEEE Xplore Abstract Record:C\:\\Users\\Mohamed Hussein\\Zotero\\storage\\5T8RBG9B\\7045992.html:text/html}
}
Downloads: 0
{"_id":"RtJJWjbnQP8iiFtzx","bibbaseid":"sharaf-torki-hussein-elsaban-realtimemultiscaleactiondetectionfrom3dskeletondata-2015","author_short":["Sharaf, A.","Torki, M.","Hussein, M. E.","El-Saban, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Real-Time Multi-scale Action Detection from 3D Skeleton Data","rights":"All rights reserved","doi":"10.1109/WACV.2015.138","abstract":"In this paper we introduce a real-time system for action detection. The system uses a small set of robust features extracted from 3D skeleton data. Features are effectively described based on the probability distribution of skeleton data. The descriptor computes a pyramid of sample covariance matrices and mean vectors to encode the relationship between the features. For handling the intra-class variations of actions, such as action temporal scale variations, the descriptor is computed using different window scales for each action. Discriminative elements of the descriptor are mined using feature selection. The system achieves accurate detection results on difficult unsegmented sequences. Experiments on MSRC-12 and G3D datasets show that the proposed system outperforms the state-of-the-art in detection accuracy with very low latency. To the best of our knowledge, we are the first to propose using multi-scale description in action detection from 3D skeleton data.","eventtitle":"2015 IEEE Winter Conference on Applications of Computer Vision","pages":"998–1005","booktitle":"2015 IEEE Winter Conference on Applications of Computer Vision","author":[{"propositions":[],"lastnames":["Sharaf"],"firstnames":["A."],"suffixes":[]},{"propositions":[],"lastnames":["Torki"],"firstnames":["M."],"suffixes":[]},{"propositions":[],"lastnames":["Hussein"],"firstnames":["M.","E."],"suffixes":[]},{"propositions":[],"lastnames":["El-Saban"],"firstnames":["M."],"suffixes":[]}],"year":"2015","keywords":"3D skeleton data, covariance matrices, Detectors, feature extraction, Feature extraction, G3D datasets, Joints, mean vectors, MSRC-12, object detection, real-time multiscale action detection, real-time system, real-time systems, Real-time systems, Three-dimensional displays, vectors, Vectors","file":"IEEE Xplore Abstract Record:C\\:\\\\Users\\\\Mohamed Hussein\\\\Zotero\\\\storage\\\\5T8RBG9B\\\\7045992.html:text/html","bibtex":"@inproceedings{sharaf_real-time_2015,\n\ttitle = {Real-Time Multi-scale Action Detection from 3D Skeleton Data},\n\trights = {All rights reserved},\n\tdoi = {10.1109/WACV.2015.138},\n\tabstract = {In this paper we introduce a real-time system for action detection. The system uses a small set of robust features extracted from 3D skeleton data. Features are effectively described based on the probability distribution of skeleton data. The descriptor computes a pyramid of sample covariance matrices and mean vectors to encode the relationship between the features. For handling the intra-class variations of actions, such as action temporal scale variations, the descriptor is computed using different window scales for each action. Discriminative elements of the descriptor are mined using feature selection. The system achieves accurate detection results on difficult unsegmented sequences. Experiments on {MSRC}-12 and G3D datasets show that the proposed system outperforms the state-of-the-art in detection accuracy with very low latency. To the best of our knowledge, we are the first to propose using multi-scale description in action detection from 3D skeleton data.},\n\teventtitle = {2015 {IEEE} Winter Conference on Applications of Computer Vision},\n\tpages = {998--1005},\n\tbooktitle = {2015 {IEEE} Winter Conference on Applications of Computer Vision},\n\tauthor = {Sharaf, A. and Torki, M. and Hussein, M. E. and El-Saban, M.},\n\tyear = {2015},\n\tkeywords = {3D skeleton data, covariance matrices, Detectors, feature extraction, Feature extraction, G3D datasets, Joints, mean vectors, {MSRC}-12, object detection, real-time multiscale action detection, real-time system, real-time systems, Real-time systems, Three-dimensional displays, vectors, Vectors},\n\tfile = {IEEE Xplore Abstract Record:C\\:\\\\Users\\\\Mohamed Hussein\\\\Zotero\\\\storage\\\\5T8RBG9B\\\\7045992.html:text/html}\n}\n\n","author_short":["Sharaf, A.","Torki, M.","Hussein, M. E.","El-Saban, M."],"bibbaseid":"sharaf-torki-hussein-elsaban-realtimemultiscaleactiondetectionfrom3dskeletondata-2015","role":"author","urls":{},"keyword":["3D skeleton data","covariance matrices","Detectors","feature extraction","Feature extraction","G3D datasets","Joints","mean vectors","MSRC-12","object detection","real-time multiscale action detection","real-time system","real-time systems","Real-time systems","Three-dimensional displays","vectors","Vectors"],"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://bibbase.org/f/2bJzYjCLapWTtM86s/mehussein-2023.bib","dataSources":["kYvtZ54PgkXqjbteW","mhdykGczo2jDicE3X","havAjNnaG4BxhYWyb"],"keywords":["3d skeleton data","covariance matrices","detectors","feature extraction","feature extraction","g3d datasets","joints","mean vectors","msrc-12","object detection","real-time multiscale action detection","real-time system","real-time systems","real-time systems","three-dimensional displays","vectors","vectors"],"search_terms":["real","time","multi","scale","action","detection","skeleton","data","sharaf","torki","hussein","el-saban"],"title":"Real-Time Multi-scale Action Detection from 3D Skeleton Data","year":2015}