Color Image Authentication and Recovery Via Adaptive Encoding

In this paper, we describe an authentication and recovery scheme for color image protection based on adaptive encoding. The image blocks are categorized based on their contents and different encoding schemes are applied according to their types. Such adaptive encoding results in better image quality and more robust image authentication. The approximations of the luminance and chromatic channels are carefully calculated, and for the purpose of reducing the data size, the differential coding is used to encode the channels with variable size according the characteristic of the block. The recovery data which represents the approximation and the detail of the image is embedded for data protection. The necessary data is well protected by using the error correcting coding (ECC) and duplication. The experimental results demonstrate that our technique is able to identify and localize image tampering, while preserving high quality for both watermarked and recovered images.