PENet+: A Lightweight Residual Transformer Framework for Efficient Image Steganalysis
About
Image steganalysis, the detection of hidden information embedded in digital images, is a core component of modern cybersecurity and digital forensics. Recent residual Transformer architectures, such as the Pixel-Difference-Convolution and Enhanced-Transformer-Network (PENet) [1], achieve strong detection accuracy, but their computational and memory demands hinder deployment in resource-constrained settings. We present PENet+, a lightweight steganalysis framework that preserves PENet's discriminative structure while substantially improving efficiency. Rather than redesigning or compressing the attention blocks, we retain PENet's self-attention topology for reproducibility and add a classifier-streamlining stage that progressively narrows the SPP-to-FC1 input channels (SPP: spatial pyramid pooling; FC1: first fully connected layer), yielding large reductions in parameters and FLOPs with negligible accuracy loss. We further refine the high-pass-filter (HPF) stem with an activation-aware mechanism that aggregates HPF responses early and selects a balanced SRM-Gabor top-K subset, and we replace PENet's backbone with a MobileNetV2-style inverted residual network. A balanced configuration with K=31 filters (16 Gabor + 15 SRM) matches or surpasses heavier settings at lower compute. Finally, we motivate PReLU from a steganalysis standpoint, arguing that preserving negative responses helps capture weak stego cues that ReLU suppresses. On a disjoint ALASKA2 JPEG QF90 protocol at 512x512 resolution (5,000 cover images for training, validation, and internal testing; a separate 19,000-cover evaluation set), PENet+ achieves up to 45.5% fewer parameters and about 97% fewer FLOPs than the re-evaluated PENet baseline, offering a computationally efficient direction for resource-constrained steganalysis. Device-level latency and power measurements remain future work.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Image Steganalysis | ALASKA2 QF90 nsF5 (test) | Accuracy (0.2 bpc)79.14 | 13 | |
| Image Steganalysis | ALASKA2 QF90 (J-UNIWARD) (test) | Accuracy (0.2 bpc)84.72 | 13 | |
| Image Steganalysis | ALASKA2 QF90 UERD (test) | Accuracy (0.2 bpc)79.68 | 13 | |
| Image Steganalysis | ALASKA2 QF90 (nsF5 @ 0.4 bpc) | Accuracy94.54 | 7 | |
| Image Steganalysis | ALASKA2 QF90 (UERD @ 0.4 bpc) | Accuracy88.86 | 7 | |
| Steganalysis | ALASKA2 QF90 256x256 nsF5 (0.2 bpc) | Accuracy68.59 | 4 | |
| Steganalysis | ALASKA2 QF90 256x256 J-UNIWARD 0.2 bpc | Accuracy76.31 | 4 | |
| Steganalysis | ALASKA2 QF90 256x256 J-UNIWARD 0.4 bpc | Accuracy86.29 | 4 | |
| Steganalysis | ALASKA2 QF90 256x256 (UERD) 0.2 bpc | Accuracy73.35 | 4 | |
| Image Steganalysis | ALASKA2 QF90 (J-UNIWARD @ 0.4 bpc) | Accuracy91.97 | 3 |