A New Approach for Information Hiding in Color Images using Adaptive Steganography and Hybrid Feature Detection with Improved PSNR and Capacity

This paper proposes a new approach for information hiding in RGB color images based on 3 new techniques 1) A new hybrid feature (line /edge /boundary /shape) detection technique combining Canny and Hough transform for bifurcating an image into edge and smooth areas 2) Two Component based Least significant bit (LSB) Substitution Technique for hiding encrypted messages in edges of images 3) An Adaptive LSB substitution technique for hiding messages to smooth areas. Firstly, the input cover image (Bitmap Image) is divided into edge and smooth areas using a new hybrid feature detection technique based on Canny Edge detection filter and Hough transform. Message bits are then embedded in the least significant bytes across edge pixels using new Two Component based LSB Substitution Technique. For the rest of the pixels across smoother areas, it utilizes an adaptive LSB .The proposed research is direct implementation of the principle that edge areas being high in contrast, color, density and frequency can tolerate more changes in their pixel values than smooth areas, so can be embedded with a large number of secret data while achieving high quality of the stego-image. Proposed technique is better than other existing techniques in comparison with PSNR, capacity and immunity to noise as shown experimentally. KeywordAdaptive LSB, Bitmap images, Canny, Hough transform, Steganography, Two component LSB insertion.