JPEG-Compatibility steganalysis detects the presence of secret message embedded in the JPEG decompressed images and estimates the embedding rate. We propose a JPEG-Compatibility steganalysis algorithm that estimates t...JPEG-Compatibility steganalysis detects the presence of secret message embedded in the JPEG decompressed images and estimates the embedding rate. We propose a JPEG-Compatibility steganalysis algorithm that estimates the embedding rate based on the difference between the stego image and its recompression based predicted co-vet image. In particular, compression artifacts and embedding changes are distinguished based on the amplitude of pixel value changes. This is done independent of the embedding positions, thus is effective for both content non-adaptive and content adaptive steganography. In addition, we also improve the recompression prediction scheme to more accurately estimate the JPEG quantization table. Experimental results show that the proposed algorithm is significantly more effective in detecting spatial ±1 steganography across a wide range of quality factors and embedding rates, when compared to the previous works.展开更多
基金This work was supported in part by National Natural Science Foundation of China，National 973 Basic Research Program of China，the Fundamental Research Funds for the Central Universities,HUST
文摘JPEG-Compatibility steganalysis detects the presence of secret message embedded in the JPEG decompressed images and estimates the embedding rate. We propose a JPEG-Compatibility steganalysis algorithm that estimates the embedding rate based on the difference between the stego image and its recompression based predicted co-vet image. In particular, compression artifacts and embedding changes are distinguished based on the amplitude of pixel value changes. This is done independent of the embedding positions, thus is effective for both content non-adaptive and content adaptive steganography. In addition, we also improve the recompression prediction scheme to more accurately estimate the JPEG quantization table. Experimental results show that the proposed algorithm is significantly more effective in detecting spatial ±1 steganography across a wide range of quality factors and embedding rates, when compared to the previous works.
