IMAGES TAMPER SELF-DETECTION AND SELF-RECOVERY USING SPIHT TECHNIQUE

摘要

This paper presents a high quality image recovery technique as a highly efficient means for image tamper detection and recovery. A region of importance (ROI) in an image is firstly compressed by a set partitioning in hierarchical tree (SPIHT) technique and then embedded into a host image. In this way, the aim of self-detection and self-recovery for a tampered image can be achieved without extra information. Higher levels of security and secrecy are reached by embedding the image data into the discrete cosine transform (DCT) frequency domain through a nested structure technique. Accordingly, a high quality recovered image is ensured. This proposal is experimentally demonstrated to provide two clear advantages, that is, 1) the aim of a high quality and highly efficient image recovery from a tempered region is reached through an ROI image compression by an SPIHT technique, and 2) a higher level of security is achieved by use of a nested embedding structure. The proposed scheme can be applied to anti-counterfeiting for archives, digital right management, protection of crime scene photos, etc.