School of Electrical and Computer Engineering, Nanfang College of Sun Yat-Sen University, Guangzhou, China.
Department of Computer Science and Information Engineering, Chang Gung University, Taoyuan, Taiwan.
PLoS One. 2020 Apr 16;15(4):e0230997. doi: 10.1371/journal.pone.0230997. eCollection 2020.
The existing tamper detection schemes for absolute moment block truncation coding (AMBTC) compressed images are able to detect the tampering. However, the marked image qualities of these schemes can be enhanced, and their authentication methods may fail to detect some special tampering. We propose a secure AMBTC tamper detection scheme that preserves high image fidelity with excellent detectability. In the proposed approach, a bit in bitmaps of AMBTC codes is sequentially toggled to generate a set of authentication codes. The one that causes the least distortion is embedded into the quantization levels with the guidance of a key-generated reference table (RT). Without the correct key, the same reference table cannot be constructed. Therefore, the proposed method is able to detect various kinds of malicious tampering, including those special tampering techniques designed for RT-based authentication schemes. The proposed method not only offers better image quality, but also provides an excellent and satisfactory detectability as compared with previous works.
现有的绝对矩块截断编码 (AMBTC) 压缩图像篡改检测方案能够检测到篡改。然而，这些方案的标记图像质量可以得到提高，并且它们的认证方法可能无法检测到某些特殊的篡改。我们提出了一种安全的 AMBTC 篡改检测方案，该方案具有出色的可检测性和高图像保真度。在提出的方法中，AMBTC 码的位图中的一位被顺序切换，以生成一组认证码。在密钥生成的参考表 (RT) 的指导下，将导致失真最小的位嵌入到量化级别中。没有正确的密钥，就无法构建相同的参考表。因此，该方法能够检测各种恶意篡改，包括针对基于 RT 的认证方案设计的特殊篡改技术。与以前的工作相比，该方法不仅提供了更好的图像质量，而且还提供了出色且令人满意的可检测性。
