arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2204.14110cs.CVcs.CY
收藏

看而不看:儿童性虐待数据集的分析管道

Seeing without Looking: Analysis Pipeline for Child Sexual Abuse Datasets

Universidade Federal de Minas Gerais(米纳斯吉拉斯联邦大学);University of Campinas(坎皮纳斯大学)

Camila Laranjeira, João Macedo, Sandra Avila, Jefersson A. dos Santos

摘要:

儿童性虐待材料(CSAM)的在线共享和查看正在快速增长,因此人类专家不再能够处理手动检查。然而,机械化中心的自动分类是一个具有挑战性的研究领域,主要是因为无法获得目标数据,而这些数据是-而且应该永远是-私人的,只有执法机构拥有。为了帮助研究人员从看不见的数据中获得见解,并安全地提供对CSAM图像的进一步理解,我们提出了一个分析模板,它超越了数据集和相应标签的统计数据。它专注于自动信号的提取,由预先训练的机器学习模型提供,例如,对象类别和色情内容检测,以及亮度和清晰度等图像指标。只提供稀疏信号的汇总统计数据,以保证受害儿童和青少年的匿名性。流水线允许通过对每个指定信号应用阈值来过滤数据,并提供子集内的此类信号的分布、信号之间的相关性以及偏差评估。我们展示了我们的建议,基于区域的注释儿童色情数据集(RCPD),在文献中的几个CSAM基准之一,由超过2000个样本之间的定期和CSAM图像,与巴西的联邦警察合作制作。虽然在几个意义上的嘈杂和有限的,我们认为,自动信号可以突出数据的整体分布,这是有价值的数据库,不能透露的重要方面。我们的目标是安全地宣传CSAM数据集的特征,鼓励研究人员加入该领域,也许其他机构也可以提供类似的基准报告。

英文摘要:

The online sharing and viewing of Child Sexual Abuse Material (CSAM) are growing fast, such that human experts can no longer handle the manual inspection. However, the automatic classification of CSAM is a challenging field of research, largely due to the inaccessibility of target data that is - and should forever be - private and in sole possession of law enforcement agencies. To aid researchers in drawing insights from unseen data and safely providing further understanding of CSAM images, we propose an analysis template that goes beyond the statistics of the dataset and respective labels. It focuses on the extraction of automatic signals, provided both by pre-trained machine learning models, e.g., object categories and pornography detection, as well as image metrics such as luminance and sharpness. Only aggregated statistics of sparse signals are provided to guarantee the anonymity of children and adolescents victimized. The pipeline allows filtering the data by applying thresholds to each specified signal and provides the distribution of such signals within the subset, correlations between signals, as well as a bias evaluation. We demonstrated our proposal on the Region-based annotated Child Pornography Dataset (RCPD), one of the few CSAM benchmarks in the literature, composed of over 2000 samples among regular and CSAM images, produced in partnership with Brazil's Federal Police. Although noisy and limited in several senses, we argue that automatic signals can highlight important aspects of the overall distribution of data, which is valuable for databases that can not be disclosed. Our goal is to safely publicize the characteristics of CSAM datasets, encouraging researchers to join the field and perhaps other institutions to provide similar reports on their benchmarks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑