Few-Shot Multi-Label Multi-Class Classification for Dark Web Image Categorization

@Article {
title = {Few-Shot Multi-Label Multi-Class Classification for Dark Web Image Categorization},
keywds = {
Multi-label image classification, Multi-class image classification, CLIP, Label empowering
}
abstract = {

Categorizing dark web image content is critical for
identifying and averting potential threats. However, this remains
a challenge due to the nature of the data, which includes multiple
co-existing domains and intra-class variations. While many methods
have been proposed to classify this image content, multi-label
multi-class classification remains underexplored. In this paper, we
propose a novel and efficient strategy for transforming a zero-shot
single-label classifier into a few-shot multi-label classifier. This
approach combines a label empowering methodology with fewshot
data. We use CLIP, a conservative learning model that uses
image-text pairs, to demonstrate the effectiveness of our strategy.
Finally, we compare our method’s performance with other multilabel
methodologies applied to CLIP and other leading multilabel
architectures.

}
isbn = {979-8-3503-3036-6},
date = {2024-04-29},
}