|
|
基于DnCNN的图像来源检验
|
Abstract:
本文基于DnCNN网络提出了一种图像来源检验方法，可以通过提取图像中的噪声残差特征(NP特征)，捕获设备特有的传感器噪声模式，并结合峰值相关能量值(PCE)的计算来判断图像的来源设备或检测篡改行为。通过在多个数据集上的实验验证，该方法能够有效区分不同型号和同一型号的不同设备拍摄的图像。研究结果表明，基于DnCNN的图像来源检验方法为数字图像取证提供了一个高效且可靠的技术手段，在公共安全、司法鉴定和媒体真实性验证等领域具有广泛的应用前景。
This paper proposes an image source verification method based on the DnCNN network, which extracts noise residual features (NP features) from images to capture device-specific sensor noise patterns. By calculating the Peak Correlation Energy (PCE), the method can determine the source device of the image or detect tampering behaviors. Experimental results on multiple datasets validate that the proposed method effectively distinguishes images captured by different devices of the same model as well as by different models. The findings indicate that the DnCNN-based image source verification method provides an efficient and reliable technological means for digital image forensics, with broad applications in public safety, forensic identification, and media authenticity verification.
| [1] | 张明旺, 肖延辉, 田华伟, 等. 图像中的设备指纹提取技术研究综述[J]. 激光与光电子学进展, 2020, 57(22): 39-48. |
| [2] | Cozzolino, D. and Verdoliva, L. (2019) Noiseprint: A CNN-Based Camera Model Fingerprint. IEEE Transactions on Information Forensics and Security, 15, 144-159. |
| [3] | 郝昕泽, 肖延辉, 田华伟, 等. 基于样本错配训练的图像PRNU噪声提纯方法[J]. 南京航空航天大学学报, 2020, 52(5): 783-791. https://doi.org/10.16356/j.1005-2615.2020.05.015 |
| [4] | Kai, Z., Zuo, W., Chen, Y., et al. (2016) Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising: IEEE Press. |
| [5] | Goljan, M. (2008) Digital Camera Identification from Images—Estimating False Acceptance Probability. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-04438-0_38 |
| [6] | Lin, X. and Li, C.T. (2017) Preprocessing Reference Sensor Pattern Noise via Spectrum Equalization. IEEE Transactions on Information Forensics & Security, 11, 126-140. https://doi.org/10.1109/TIFS.2015.2478748 |
