High-Quality Jpeg Compression Using LDN Comparison and Quantization Noise Analysis

To identify whether an image has been JPEG compressed is an important issue in forensic practice.. In this paper, we provide a novel quantization noise-based solution to reveal the traces of JPEG compression with preprocessing system for extracting LDN (local directional number pattern) before compression& during decompression LDN get compared. What LDN does is generates 6 digit binary code for each pixel and stores separately during compression. LDN value of image before compression and after compression is compared to get significant result. We analytically derive that a decompressed JPEG image has a lower variance of forward quantization noise than its uncompressed counterpart. With the conclusion, we develop a simple yet very effective detection algorithm to identify decompressed JPEG images. We also demonstrate that the proposed method is robust to small image size and chroma sub sampling. The proposed algorithm can be applied in some practical applications, such as Internet image classification and forgery detection.