Researchers have developed CrosInv, an efficient cross-scale invertible hiding network designed for concealing image-level messages within cover images. This new network addresses limitations of existing methods by employing cross-scale and spatial-frequency collaborative features, enhancing nonlinear representation capabilities. CrosInv utilizes a cross-scale invertible module with pixel shuffle and Haar wavelet transformations for scale integration, alongside a non-invertible cross dense module to boost nonlinearity. Experimental results demonstrate the effectiveness and superiority of the CrosInv network. AI
IMPACT Introduces a novel network architecture that could improve data hiding techniques in digital images.
RANK_REASON The cluster contains a research paper detailing a new network architecture for image hiding. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
