Attention-Guided Network Pruning Compresses Image Steganalysis Models While Preserving Detection Performance
Researchers designed an attention-guided pruning framework named SEAP that shrinks inverted bottleneck steganalysis architectures while preserving detection accuracy.

