CALPA-NET: Channel-pruning-assisted Deep Residual Network for Steganalysis of Digital Images

TL;DR

CALPA-NET is a novel approach that prunes deep residual networks for digital image steganalysis, significantly reducing model size and computational cost while maintaining high detection performance.

Contribution

This paper introduces CALPA-NET, a data-driven network pruning method that creates compact, efficient steganalytic models from over-parameterized deep networks, improving deployability.

Findings

Achieves comparable steganalysis performance with less than 2% of original parameters.

Reduces FLOPs by approximately two-thirds compared to baseline models.

Demonstrates superior transferability and scalability across multiple datasets.

Abstract

Over the past few years, detection performance improvements of deep-learning based steganalyzers have been usually achieved through structure expansion. However, excessive expanded structure results in huge computational cost, storage overheads, and consequently difficulty in training and deployment. In this paper we propose CALPA-NET, a ChAnneL-Pruning-Assisted deep residual network architecture search approach to shrink the network structure of existing vast, over-parameterized deep-learning based steganalyzers. We observe that the broad inverted-pyramid structure of existing deep-learning based steganalyzers might contradict the well-established model diversity oriented philosophy, and therefore is not suitable for steganalysis. Then a hybrid criterion combined with two network pruning schemes is introduced to adaptively shrink every involved convolutional layer in a data-driven manner. The resulting network architecture presents a slender bottleneck-like structure. We have conducted extensive experiments on BOSSBase+BOWS2 dataset, more diverse ALASKA dataset and even a large-scale subset extracted from ImageNet CLS-LOC dataset. The experimental results show that the model structure generated by our proposed CALPA-NET can achieve comparative performance with less than two percent of parameters and about one third FLOPs compared to the original steganalytic model. The new model possesses even better adaptivity, transferability, and scalability.