Steganalysis attacks resilient, imperceptible, blind, and keyed image hiding in ultra HD 4K video

Abstract

ABSTRACT A novel data hiding approach is being proposed in this article, camouflaging secret data within a very high-quality video object. The proposed algorithm utilizes Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) to hide image data, preprocessed using Arnold Transform. The novelty of this approach lies in the combination of five factors, which are – use of very high-resolution Ultra HD 4 K video as a cover object; Use of DCT coefficient in Scene Change Detection to identify cover frame; Scrambling the secret image by Arnold Transformation; Use of DCT-DWT in embedding and extraction process to that hide data in the blue plane of the RGB frame; and lastly, use of three different keys for data hiding to secure the proposed method against attacks, by adhering the Kerckhoff’s principle of a cryptosystem. Several quality metrics like Structural Content, Peak Signal-to-Noise Ratio, Normalized Cross-Correlation, Average Difference, Maximum Difference, and Normalized Absolute Error have been calculated to validate the imperceptibility. This method has also been validated against different steganalysis attacks like random cropping, rotating, resizing, noise addition, filtering, histogram attack, and lastly, compared with existing data hiding methods to prove superiority.