Dynamic video steganography using LBP on CIELAB based K-means clustering

A video is a collection of still images called frames and audio. This paper presents an approach to video steganography for hiding secret data using Local Binary Pattern (LBP) in the desired segment formed by the k-means clustering of the desired frame. For embedding and extracting the message LBP is used. LBP, using the pattern classification characteristics, alters values of cover image pixels, so that the modification qualifies the message requirements and aids the extraction. As pure LBP for steganography can be detectable and messages would be retrievable, it is secured by enabling Choice-selection of location for hiding the message in the desired segment of a frame. Segmentation is performed using k-means clustering as it provides better accuracy for huge data sets. As each frame has huge number of pixels and a pixel furthermore has color components, the k-means clustering serves the purpose also concerning the speed. For the k-means clustering CIELAB (Informally called Lab) color model is used as the Luminance component is separable and segmentation is done only on basis of chromatic components of Lab. The whole approach is dynamic because as it accepts selection of Frame, Cluster of Choice Number, in which Message is to be hidden or extracted from.