DrLS: Distortion‐Resistant Lossless Steganography via Colour Depth Interpolation

ABSTRACT The lossless data steganography is to hide a certain amount of information into a container image. Previous lossless steganography methods fail to strike a balance between capacity, imperceptibility, accuracy, and robustness, commonly vulnerable to distortion on container images. This paper introduces DrLS, a distortion‐resistant lossless steganography, enabling secure hiding and lossless revealing of arbitrary messages within visually realistic container images. DrLS innovates by hiding the bitstream through the process of colour enhancement, concretely, a bit‐depth resampling problem. For the first time, a swim‐transformer‐based residual block (TransRB) is implemented inside the flow model to boost performance, along with Conditional Modulation over Flow‐based models (CMF). Extensive experiments demonstrate that DrLS achieves unparallelled robustness and capacity under pristine and distorted conditions, setting a new benchmark for data‐hiding technologies. DrLS's contributions promise substantial impacts on secure communication and digital media protection, heralding a new era of steganography solutions.