Forensic Science International

2016

DOI: 10.1016/j.forsciint.2016.09.010

|View full text |Cite

|

Sign up to set email alerts

|

Pornography classification: The hidden clues in video space–time

¹

,

²

,

Mauricio Pérez

³

et al.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...

Methods1

Pooling Late Fusion Classification1

A Elsagate Dataset1

II Related Work1

Citation Types

Supporting

0

Mentioning

58

Contrasting

0

Unclassified

1

Year Published

2018

2018

2024

2024

Publication Types

Select...

Book5

Other2

Article1

Relationship

Self Cite2

Independent6

Authors

Journals

Cited by 63 publications

(59 citation statements)

References 19 publications

Supporting

0

Mentioning

58

Contrasting

0

Unclassified

1

Order By: Relevance

“…These features have been identified as important in related implementations of the pipeline Moreira et al (2019). Interpretable features can be extracted from the pipeline using the learned codebook, as described by Moreira et al (2016).…”

Section: Methodsmentioning

confidence: 99%

Getting the subtext without the text: Scalable multimodal sentiment classification from visual and acoustic modalities

et al. 2018

Proceedings of Grand Challenge and Workshop on Human Multimodal Language (Challenge-Hml)

Self Cite

View full text Add to dashboard Cite

In the last decade, video blogs (vlogs) have become an extremely popular method through which people express sentiment. The ubiquitousness of these videos has increased the importance of multimodal fusion models, which incorporate video and audio features with traditional text features for automatic sentiment detection. Multimodal fusion offers a unique opportunity to build models that learn from the full depth of expression available to human viewers. In the detection of sentiment in these videos, acoustic and video features provide clarity to otherwise ambiguous transcripts. In this paper, we present a multimodal fusion model that exclusively uses high-level video and audio features to analyze spoken sentences for sentiment. We discard traditional transcription features in order to minimize human intervention and to maximize the deployability of our model on at-scale real-world data. We select high-level features for our model that have been successful in nonaffect domains in order to test their generalizability in the sentiment detection domain. We train and test our model on the newly released CMU Multimodal Opinion Sentiment and Emotion Intensity (CMU-MOSEI) dataset, obtaining an F 1 score of 0.8049 on the validation set and an F 1 score of 0.6325 on the held-out challenge test set. Figure 1: A blindspot in multimodal sentiment analysis is the inclusion of human-transcriptions of spoken sentiment, which limits model applicability. We address this by using only prosodic and visual features for sentiment classification.

“…These features have been identified as important in related implementations of the pipeline Moreira et al (2019). Interpretable features can be extracted from the pipeline using the learned codebook, as described by Moreira et al (2016).…”

Section: Methodsmentioning

confidence: 99%

Getting the subtext without the text: Scalable multimodal sentiment classification from visual and acoustic modalities

et al. 2018

Proceedings of Grand Challenge and Workshop on Human Multimodal Language (Challenge-Hml)

Self Cite

View full text Add to dashboard Cite

In the last decade, video blogs (vlogs) have become an extremely popular method through which people express sentiment. The ubiquitousness of these videos has increased the importance of multimodal fusion models, which incorporate video and audio features with traditional text features for automatic sentiment detection. Multimodal fusion offers a unique opportunity to build models that learn from the full depth of expression available to human viewers. In the detection of sentiment in these videos, acoustic and video features provide clarity to otherwise ambiguous transcripts. In this paper, we present a multimodal fusion model that exclusively uses high-level video and audio features to analyze spoken sentences for sentiment. We discard traditional transcription features in order to minimize human intervention and to maximize the deployability of our model on at-scale real-world data. We select high-level features for our model that have been successful in nonaffect domains in order to test their generalizability in the sentiment detection domain. We train and test our model on the newly released CMU Multimodal Opinion Sentiment and Emotion Intensity (CMU-MOSEI) dataset, obtaining an F 1 score of 0.8049 on the validation set and an F 1 score of 0.6325 on the held-out challenge test set. Figure 1: A blindspot in multimodal sentiment analysis is the inclusion of human-transcriptions of spoken sentiment, which limits model applicability. We address this by using only prosodic and visual features for sentiment classification.

“…2) Motion Information: Incorporating motion information into local descriptors and deep neural networks leads to more effective sensitive videos classifiers [12], [17], [24]. In this paper, we aim to develop deep learning-based approaches for automatically extracting discriminative space-temporal information for filtering Elsagate content, with a good compromise between effectiveness and efficiency.…”

Section: Pooling Late Fusion Classificationmentioning

confidence: 99%

“…It comprises 285 hours (1,028,106 seconds) of 1,396 Elsagate and 1,898 non-sensitive videos. To put the number in perspective, the largest sensitive video dataset (Pornography-2k dataset [24]) contains 140 hours. It is worth mentioning the Elsagate dataset is composed of cartoons only.…”

Section: A Elsagate Datasetmentioning

confidence: 99%

Combating the Elsagate Phenomenon: Deep Learning Architectures for Disturbing Cartoons

¹

,

²

,

³

2019

2019 7th International Workshop on Biometrics and Forensics (IWBF)

Self Cite

View full text Add to dashboard Cite

Watching cartoons can be useful for children's intellectual, social and emotional development. However, the most popular video sharing platform today provides many videos with Elsagate content. Elsagate is a phenomenon that depicts childhood characters in disturbing circumstances (e.g., gore, toilet humor, drinking urine, stealing). Even with this threat easily available for children, there is no work in the literature addressing the problem. As the first to explore disturbing content in cartoons, we proceed from the most recent pornography detection literature applying deep convolutional neural networks combined with static and motion information of the video. Our solution is compatible with mobile platforms and achieved 92.6% of accuracy. Our goal is not only to introduce the first solution but also to bring up the discussion around Elsagate.

“…The BoF model was used in [20] where authors presented to use a multi-instance modeling scheme based on spatial pyramid partitions (SPP) to transfer the target problem (pornography detection) into MIL problem. Temporal Robust Features (TRoF) were for the first time employed in the task of pornography detection videos by the work in [7]. It was proposed to aggregate the TRoF features into mid-level features by using FV (Fisher vector), a new variation of BoVW (bag of visual words) model.…”

Section: II Related Workmentioning

confidence: 99%

“…Although there are several attempts in the literature in order to address the problem of nudity detection, there is variation in the definition of the word "Nudity" in our work with those previous scholarly works [3], [7]- [11]. For instance, a women wearing a bikini is considered a regular content in USA or Europe, but it is defined as adult content in Malaysia (and even other countries like Indonesia, or Brunei).…”

Section: Introductionmentioning

confidence: 99%

Automated Nudity Recognition using Very Deep Residual Learning Network

¹

,

²

,

³

et al. 2019

View full text Add to dashboard Cite

The exponentially growing number of pornographic material has brought many challenges to the modern daily life, particularly where children and minors have unlimited access to the internet. In Malaysia, all local and foreign films should obtain the suitability approval before distribution or public viewing, and this process of screening visual contents of all the TV channels imposes a huge censorship cost to the service providers such as Unifi TV. To leverage this issue, this paper proposes to use an emerging model of Deep Learning (DL) techniques called Residual Learning Convolutional Neural Networks (ResNet), in order to automate the process of nudity detection in visual contents. The pre-trained ResNet model, with hundred and one layers, was utilized to perform transfer learning and solve a new binary classification problem of nudity versus non-nudity. The performance of the proposed model is evaluated based on a newly created dataset comprising more than 4k samples of nudity and non-nudity images. After conducting experiments on the nudity dataset, the deep learning method succeeded to achieve the best performance of 70.42% in term of F-score, 84.04% in term of accuracy, and 93.72% in term of AUC .

scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.

Contact Info

hi@scite.ai

10624 S. Eastern Ave., Ste. A-614

Henderson, NV 89052, USA

Product

Browser Extension Assistant by scite Citation Statement Search Reference Check Visualizations Dashboards Explore Journals Explore Organizations Explore Funders Embedding Badge Embedding Citation Search Pricing

Resources

Blog Help & FAQ Accessibility Statement API Terms For Universities & Governments For Researchers For Publishers For Corporate, Pharma & Enterprise Author Marketing Become an Affiliate Get an organization trial or quote scite Data & Services

About

News & Press Careers Read our Paper Coverage

Blog Terms and Conditions API Terms Privacy Policy Contact Cookie Preferences Do Not Sell or Share My Personal Information

Copyright © 2024 scite LLC. All rights reserved.

Made with 💙 for researchers

Part of the Research Solutions Family.