Illegal Activity Categorisation in DarkNet Based on Image Classification Using CREIC Method

Fidalgo, Eduardo; Alegre, Enrique; González-Castro, V́ıctor; Fernández-Robles, Laura

doi:10.1007/978-3-319-67180-2_58

Cited by 14 publications

(17 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In this work we extended the TOr Image Categories (TOIC) dataset proposed by [16]. Figure 2 shows an example of an image from TOIC per class.…”

Section: Datasetmentioning

confidence: 99%

“…The latter value of visual words, i.e. 2048, was chosen following the settings where TOIC dataset was published [16]. We used a hard assignment approach [31] to represent the images with feature vectors.…”

Section: Implementation Detailsmentioning

confidence: 99%

“…Fidalgo et al proposed Compass Radius Estimation for Improved Image Classification (CREIC) [15], a method based on Edge-SIFT descriptors to estimate an optimized radius value for the compass operator to compute Edge-SIFT descriptors. Later, they used CREIC method to characterize images related to illegal activities in the Tor network, presenting a new dataset named TOIC (TOr Image Categories) [16], which contains images from Tor domains labelled as drugs, weapons, money, personal ID or credit cards, depending on the category of its corresponding domain in the DUTA (Darknet Usage Text Addresses) dataset [2,17]. Later, they proposed Au-toBlur and two variants of Semantic Attention Region Filtering (SARF) methods [18], which create richer dictionaries to boost the performance of the BoVW model, which also used dense SIFT to extract low-level features from the image.…”

Section: Introductionmentioning

confidence: 99%

“…this work, we have used the Bag of Visual Words (BoVW) model to classify images crawled from domains of the the Tor network into the illegal activity according to the domain they belong to. Specifically, we have extended the TOIC (TOr Image Categories) dataset[16] with 15 new images extracted from the crawling of 500 new Tor do-…”

mentioning

confidence: 99%

See 3 more Smart Citations

Color Sift Descriptors to Categorize Illegal Activities in Images of Onion Domains

Matilla¹,

González-Castro²,

Fernández-Robles³

et al. 2020

Actas De Las XXXIX Jornadas De Automática, Badajoz, 5-7 De Septiembre De 2018

Self Cite

View full text Add to dashboard Cite

Dark Web, i.e. the portion of the Web whose content is not indexed either accessible by standard web browsers, comprises several darknets. The Onion Router (Tor) is the most famous one, thanks to the anonymity provided to its users, and it results in the creation of domains, or hidden services, which hosts illegal activities. In this work, we explored the possibility of identifying illegal domains on Tor darknet based on its visual content. After crawling and filtering the images of 500 hidden services, we sorted them into five different illegal categories, and we trained a classifier using the Bag of Visual Words (BoVW) model. In this model, SIFT (Scale Invariant Feature Transform) or dense SIFT were used as the descriptors of the images patches to compute the visual words of the BoVW model. However, SIFT only works with gray-scale images; thus the information given by color in an image is not retrieved. To overcome this drawback, in this work we implemented and assessed the performance of three different variants of SIFT descriptors that can be used in color images, namely HSV-SIFT, RGB-SIFT and Opponent-SIFT, in the BoVW model for image classification. The obtained results showed the usefulness of using color-SIFT descriptors instead of SIFT, whereas in our experiments the latter achieved an accuracy of 57.52%, the HSV-SIFT descriptor achieved an accuracy up to 59.44%.

show abstract

“…In this work we extended the TOr Image Categories (TOIC) dataset proposed by [16]. Figure 2 shows an example of an image from TOIC per class.…”

Section: Datasetmentioning

confidence: 99%

Section: Implementation Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Color Sift Descriptors to Categorize Illegal Activities in Images of Onion Domains

Matilla¹,

González-Castro²,

Fernández-Robles³

et al. 2020

Actas De Las XXXIX Jornadas De Automática, Badajoz, 5-7 De Septiembre De 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…For this purpose, we will make use of two components: text detection using a connectionist text proposal network algorithm and text recognition using an end-to-end trainable neural network. Then, in order to test the performance of the proposed pipeline, we considered a subset of 100 images from the TOIC dataset [9], which contains five different categories of images related to different illegal activities from the Tor network ( Figure 1). We set a baseline methodology by fixing the elements of the Text Spotting pipeline, which will allow future researches to compare their obtained results.…”

Section: Introductionmentioning

confidence: 99%

Detecting Textual Information in Images from Onion Domains Using Text Spotting

Blanco¹,

Fidalgo²,

Alegre³

et al. 2020

Actas De Las XXXIX Jornadas De Automática, Badajoz, 5-7 De Septiembre De 2018

Self Cite

View full text Add to dashboard Cite

Due to the efforts of different authorities in the fight against illegal activities in the Tor networks, the traders have developed new ways of circumventing the monitoring tools used to obtain evidence of said activities. In particular, embedding textual content into graphical objects avoids that text analysis, using Natural Language Processing (NLP) algorithms, can be used for watching such onion web contents. In this paper, we present a Text Spotting framework dedicated to detecting and recognizing textual information within images hosted in onion domains. We found that the Connectionist Text Proposal Network and Convolutional Recurrent Neural Network achieve 0.57 F-Measure when running the combined pipeline on a subset of 100 images labeled manually obtained from TOIC dataset. We also identified the parameters that have a critical influence on the Text Spotting results. The proposed technique might support tools to help the authorities in detecting these activities.

show abstract

Making Sense of Darknet Markets: Automatic Inference of Semantic Classifications from Unconventional Multimedia Datasets

Berman

Paul

2019

HCI for Cybersecurity, Privacy and Trust

View full text Add to dashboard Cite

Illegal Activity Categorisation in DarkNet Based on Image Classification Using CREIC Method

Cited by 14 publications

References 17 publications

Color Sift Descriptors to Categorize Illegal Activities in Images of Onion Domains

Color Sift Descriptors to Categorize Illegal Activities in Images of Onion Domains

Detecting Textual Information in Images from Onion Domains Using Text Spotting

Making Sense of Darknet Markets: Automatic Inference of Semantic Classifications from Unconventional Multimedia Datasets

Contact Info

Product

Resources

About