Adaptive Steganography via Image Complexity Analysis using 3D Color Texture Feature

• Published in: 2018 3rd International Innovative Applications of Computational Intelligence on Power, Energy and Controls with their Impact on Humanity (CIPECH) pp. 1 - 5
• Main Authors: Grover, Ridhima, Yadav, Dinesh Kumar, Chauhan, D.K., Kamya, Suraj
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2018
• Subjects: adaptive steganography; color texture feature; complexity analysis; Complexity theory; CTF; Data mining; DCT; Discrete cosine transforms; Feature extraction; Frequency-domain analysis; GLCM texture; Image color analysis; Three-dimensional displays
• DOI: 10.1109/CIPECH.2018.8724225

Human vision system is generally a subjective perception which varies as per individuals. Complexity of an image plays significant role while securing data in to it. In this paper a new steganography approach is presented which uses combined 3D Color Texture Feature (CTF) to identify complex regions of image for data hiding so that visual attack to detect hidden message becomes quite challenging. Frequency domain is used to hide the data in these selected complex regions via Discrete Cosine Transform (DCT). These kinds of areas are originally noisy and extracting extra information is very hard. Each image has different complexity levels and spatial regions, and since data hiding is directly dependent on it, so the steganography system becomes adaptive. The result shows that proposed adaptive method provides secure message hiding while maintaining imperceptibility quality and high embedding capacity. Final cover images maintains PSNR value of above 50. Embedding capacity is around 2 times higher in comparison to similar algorithm which uses Gray Level Co-occurrence Matrices (GLCM) feature to identify complex regions of images for data hiding.