Derivative-based audio steganalysis

Abstract

This article presents a second-order derivative-based audio steganalysis. First, Mel-cepstrum coefficients and Markov transition features from the second-order derivative of the audio signal are extracted; a support vector machine is then applied to the features for discovering the existence of hidden data in digital audio streams. Also, the relation between audio signal complexity and steganography detection accuracy, which is an issue relevant to audio steganalysis performance evaluation but so far has not been explored, is analyzed experimentally. Results demonstrate that, in comparison with a recently proposed signal stream-based Mel-cepstrum method, the second-order derivative-based audio steganalysis method gains a considerable advantage under all categories of signal complexity-especially for audio streams with high signal complexity, which are generally the most challenging for steganalysis-and thereby significantly improves the state of the art in audio steganalysis.

"A MultiAlgorithm, High Reliability, Extensible Steganalyzer using Services Oriented Architecture"

“…(1) All the three classification techniques for our improved 2-D M el-Cepstrum implementation outperform the results reported by Liu [37]. NN_Mix which uses a training set obtained by mixing the data sets from the three embedding techniques.…”

“…The complexity service is implemented for the files that have an extension of .wav. The choice of two different service implementations of .wav files is motivated by the reported results in [28,37] that as the complexity of the .wav audio files increases, a different technique can be used to obtain better stego detection accuracy results. The M el-Cepstrum service provided by our steganalyzer is based on M el-frequency cepstrum technique in conjunction with second order derivative, which is widely used in image processing to detect isolated points, edges, etc [38].…”

“…The second column in the table shows the average testing accuracy results reported by Liu for a data set of 19380 using SVM s classifiers [37]. The data set used by Liu has similar characteristics (mono wav files 44.1 kHz with 16 bit PCM quantization.…”

“…We present the results on this improved classifier in the next section. [37,40]. Accordingly, we revised our initial architecture to remove the M arkov service as it becomes non-beneficial to include it.…”

A Novel AMR-WB Speech Steganography Based on Diameter-Neighbor Codebook Partition

“…For example, when = 48, twice the embedding capacity of CNV-based method may be obtained with a decrease of only about 2% in speech quality and much the same undetectability. Moreover, both the quality of stego speech and the security of defending against statistical steganalysis [17,18] are better than the recent NID-based speech steganography.…”