2021
Santos, F.; Durães, D.; Marcondes, F. S.; Lange, S.; Machado, J.; Novais, P.
Efficient Violence Detection Using Transfer Learning Proceedings Article
Em: F., Duraes D. El Bolock A. De La Prieta (Ed.): pp. 65-75, Springer Science and Business Media Deutschland GmbH, 2021, ISSN: 18650929, (cited By 4; Conference of International Conference on Practical Applications of Agents and Multi-Agent Systems, PAAMS 2021 ; Conference Date: 6 October 2021 Through 9 October 2021; Conference Code:266119).
Resumo | Links | BibTeX | Etiquetas: Action recognition; Activity recognition; Deep learning; Fine tuning; General patterns; Learn+; State of the art; Transfer learning; Violence detections; Violent behavior, Deep learning, Human computer interaction; Pattern recognition; Security systems
@inproceedings{Santos202165,
title = {Efficient Violence Detection Using Transfer Learning},
author = {F. Santos and D. Durães and F. S. Marcondes and S. Lange and J. Machado and P. Novais},
editor = {Duraes D. El Bolock A. De La Prieta F.},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85116505876&doi=10.1007%2f978-3-030-85710-3_6&partnerID=40&md5=c229aac41b0bc0f030eb19d748792d78},
doi = {10.1007/978-3-030-85710-3_6},
issn = {18650929},
year = {2021},
date = {2021-01-01},
journal = {Communications in Computer and Information Science},
volume = {1472 CCIS},
pages = {65-75},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {In recent years several applications, namely in surveillance, human-computer interaction and video recovery based on its content has studied the detection and recognition of violence [22]. The purpose of violence detection is to automatically and effectively determine whether or not violence occurs in a short time. So, it is a crucial area since it will automatically enable the necessary means to stop the violence. To quickly solve this problem, we used models trained to solve general activity recognition problems such as Kinetics-400 to learn to extract general patterns that are very important to detect violent behaviour in videos. Our approach consists of using a state of the art pre-trained model in general activity recognition tasks (e.g. Kinetics-400) and then fine-tuning it to violence detection. We applied this approach in two violence datasets and achieved state-of-the-art results using only four input frames. © 2021, Springer Nature Switzerland AG.},
note = {cited By 4; Conference of International Conference on Practical Applications of Agents and Multi-Agent Systems, PAAMS 2021 ; Conference Date: 6 October 2021 Through 9 October 2021; Conference Code:266119},
keywords = {Action recognition; Activity recognition; Deep learning; Fine tuning; General patterns; Learn+; State of the art; Transfer learning; Violence detections; Violent behavior, Deep learning, Human computer interaction; Pattern recognition; Security systems},
pubstate = {published},
tppubtype = {inproceedings}
}
In recent years several applications, namely in surveillance, human-computer interaction and video recovery based on its content has studied the detection and recognition of violence [22]. The purpose of violence detection is to automatically and effectively determine whether or not violence occurs in a short time. So, it is a crucial area since it will automatically enable the necessary means to stop the violence. To quickly solve this problem, we used models trained to solve general activity recognition problems such as Kinetics-400 to learn to extract general patterns that are very important to detect violent behaviour in videos. Our approach consists of using a state of the art pre-trained model in general activity recognition tasks (e.g. Kinetics-400) and then fine-tuning it to violence detection. We applied this approach in two violence datasets and achieved state-of-the-art results using only four input frames. © 2021, Springer Nature Switzerland AG.
Santos, F.; Durães, D.; Marcondes, F.; Gomes, M.; Gonçalves, F.; Fonseca, J.; Wingbermuehle, J.; Machado, J.; Novais, P.
Modelling a Deep Learning Framework for Recognition of Human Actions on Video Proceedings Article
Em: A., Dzemyda G. Adeli H. Rocha (Ed.): pp. 104-112, Springer Science and Business Media Deutschland GmbH, 2021, ISSN: 21945357, (cited By 3; Conference of World Conference on Information Systems and Technologies, WorldCIST 2021 ; Conference Date: 1 April 2021 Through 2 April 2021; Conference Code:256979).
Resumo | Links | BibTeX | Etiquetas: Action recognition; Discriminative features; High-performance hardware; Human activities; Human-action recognition; Intelligent solutions; Learning frameworks; Learning models, Deep learning, Information systems; Information use; Learning systems
@inproceedings{Santos2021104,
title = {Modelling a Deep Learning Framework for Recognition of Human Actions on Video},
author = {F. Santos and D. Durães and F. Marcondes and M. Gomes and F. Gonçalves and J. Fonseca and J. Wingbermuehle and J. Machado and P. Novais},
editor = {Dzemyda G. Adeli H. Rocha A.},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85105949815&doi=10.1007%2f978-3-030-72657-7_10&partnerID=40&md5=63526ae46868c827835600c4dba3711b},
doi = {10.1007/978-3-030-72657-7_10},
issn = {21945357},
year = {2021},
date = {2021-01-01},
journal = {Advances in Intelligent Systems and Computing},
volume = {1365 AIST},
pages = {104-112},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {In Human action recognition, the identification of actions is a system that can detect human activities. The types of human activity are classified into four different categories, depending on the complexity of the steps and the number of body parts involved in the action, namely gestures, actions, interactions, and activities [1]. It is challenging for video Human action recognition to capture useful and discriminative features because of the human body's variations. To obtain Intelligent Solutions for action recognition, it is necessary to training models to recognize which action is performed by a person. This paper conducted an experience on Human action recognition compare several deep learning models with a small dataset. The main goal is to obtain the same or better results than the literature, which apply a bigger dataset with the necessity of high-performance hardware. Our analysis provides a roadmap to reach the training, classification, and validation of each model. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.},
note = {cited By 3; Conference of World Conference on Information Systems and Technologies, WorldCIST 2021 ; Conference Date: 1 April 2021 Through 2 April 2021; Conference Code:256979},
keywords = {Action recognition; Discriminative features; High-performance hardware; Human activities; Human-action recognition; Intelligent solutions; Learning frameworks; Learning models, Deep learning, Information systems; Information use; Learning systems},
pubstate = {published},
tppubtype = {inproceedings}
}
In Human action recognition, the identification of actions is a system that can detect human activities. The types of human activity are classified into four different categories, depending on the complexity of the steps and the number of body parts involved in the action, namely gestures, actions, interactions, and activities [1]. It is challenging for video Human action recognition to capture useful and discriminative features because of the human body's variations. To obtain Intelligent Solutions for action recognition, it is necessary to training models to recognize which action is performed by a person. This paper conducted an experience on Human action recognition compare several deep learning models with a small dataset. The main goal is to obtain the same or better results than the literature, which apply a bigger dataset with the necessity of high-performance hardware. Our analysis provides a roadmap to reach the training, classification, and validation of each model. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.