First Quantization Coefficient Extraction from Double Compressed JPEG Images

Abstract

In the forensics domain can be useful to recover image history, and in particular whether or not it has been doubly compressed. Clarify this point allows to assess if, in addition to the compression at the time of shooting, the picture was decompressed and then resaved. This is not a clear indication of forgery, but it can justify further investigations. In this paper we propose a novel technique able to retrieve the coefficients of the first compression in a double compressed JPEG image when the second compression is lighter than the first one. The proposed approach exploits the effects of successive quantizations followed by dequantization to recover the original compression parameters. Experimental results and comparisons with a state of the art method confirm the effectiveness of the proposed approach.

Estimating JPEG compression history of bitmaps based on factor histogram

“…Recently, Lin et al [9] further improved Ye et al's method by classifying the patterns of individual frequencies into four classes, and then extracting the quantization step adaptively from the energy density spectrum (EDS) of DCT coefficient histogram and the Fourier transform of EDS. In addition, there are also some works aiming to detect doubly JPEG compressed images and estimate quantization parameters of the primary JPEG compression [14][15][16][17][18][19][20][21]. These works typically assumed that the quantized DCT coefficients can be directly decoded from the input JPEG file, or the quantization table of the last (i.e., secondary) JPEG compression is known.…”