Publications
2021
1.
Gangwar, Abhishek; Castro, Víctor González; Alegre, Enrique; Fidalgo, Eduardo
AttM-CNN: Attention and metric learning based CNN for pornography, age and Child Sexual Abuse (CSA) Detection in images Artículo de revista
En: Neurocomputing, vol. 445, pp. 81–104, 2021, (Publisher: Elsevier).
Resumen | Enlaces | BibTeX | Etiquetas: age-group detection, CNN, convolutional neural network, CSA detection, metric learning, pornography detection, visual attention
@article{gangwar_attm-cnn_2021,
title = {AttM-CNN: Attention and metric learning based CNN for pornography, age and Child Sexual Abuse (CSA) Detection in images},
author = {Abhishek Gangwar and Víctor González Castro and Enrique Alegre and Eduardo Fidalgo},
url = {https://www.sciencedirect.com/science/article/pii/S092523122100312X},
year = {2021},
date = {2021-01-01},
journal = {Neurocomputing},
volume = {445},
pages = {81–104},
abstract = {This paper proposes AttM-CNN, a deep learning model for detecting Child Sexual Abuse (CSA) material by combining pornographic content detection and age-group classification. Two new datasets, Pornographic-2M and Juvenile-80k, are introduced for training. The model outperforms state-of-the-art methods and improves CSA detection accuracy over forensic tools, aiding law enforcement.},
note = {Publisher: Elsevier},
keywords = {age-group detection, CNN, convolutional neural network, CSA detection, metric learning, pornography detection, visual attention},
pubstate = {published},
tppubtype = {article}
}
This paper proposes AttM-CNN, a deep learning model for detecting Child Sexual Abuse (CSA) material by combining pornographic content detection and age-group classification. Two new datasets, Pornographic-2M and Juvenile-80k, are introduced for training. The model outperforms state-of-the-art methods and improves CSA detection accuracy over forensic tools, aiding law enforcement.