High Capacity Multi-level Image Steganography Model Using KNN Classifier

Abstract

In the digital world to maintain confidentiality, integrity and authenticity of the web, audio or video communication Steganography is the mostly used technique. The major characteristics of information hiding scheme is to increase the payload of secret data with less visual distortion and better secrecy. For information hiding, identifying the suitable locality to embed the key message but with the smallest amount of deformation of the cover medium remains a troublesome matter. To achieve this goal, we propose a unique information hiding method which results in enhanced performance. The extracted attributes of the image is given as input to the KNN classifier. It classifies the block and determines the hiding capacity of each block. To provide better hiding the secret message is encoded using Adaptive Huffman encoding technique. Some evaluating metric both Qualitative and Quantitative like Structural similarity index (SSIM), PSNR and MSE are calculated to compare the performance of the proposed method with the existing information hiding approaches. The proposed scheme results in better performance with respect to capacity of hiding with less noise.