Classifying Suspicious Content in Tor Darknet

TL;DR

This paper presents a novel method combining saliency maps with Bag of Visual Words to automatically classify suspicious images in the Tor Darknet, significantly improving accuracy over existing CNN and BoVW methods.

Contribution

It introduces Semantic Attention Keypoint Filtering (SAKF), a new pixel-level feature filtering technique for better classification of Darknet images.

Findings

Achieved 87.98% accuracy on a custom Tor dataset.

SAKF outperforms CNN-based features and traditional BoVW methods.

Demonstrates effectiveness of saliency-guided feature filtering in Darknet image classification.

Abstract

One of the tasks of law enforcement agencies is to find evidence of criminal activity in the Darknet. However, visiting thousands of domains to locate visual information containing illegal acts manually requires a considerable amount of time and resources. Furthermore, the background of the images can pose a challenge when performing classification. To solve this problem, in this paper, we explore the automatic classification Tor Darknet images using Semantic Attention Keypoint Filtering, a strategy that filters non-significant features at a pixel level that do not belong to the object of interest, by combining saliency maps with Bag of Visual Words (BoVW). We evaluated SAKF on a custom Tor image dataset against CNN features: MobileNet v1 and Resnet50, and BoVW using dense SIFT descriptors, achieving a result of 87.98% accuracy and outperforming all other approaches.