Research On Steganography And Steganalysis Technologies For ILBC Speech Streams

Encryption technology may not effectively meet the user's protection of private information,while steganography can conceal the existence of private information,and it has become an important research branch in the field of information security.With the rapid development of VOIP(Voice over Internet Protocol),secret communication with low bit-rate compressed speech streams as hidden carrier has gradually become one of the key issues in steganography research.However,the existing speech steganography technology has low embedding capacity and cannot resist the detection of corresponding steganalysis technology.On the other hand,in order to prevent the transitional use of steganography,steganalysis technology can effectively supervise and detect digital carriers,and it is very important to prevent the loss of confidential information and monitor illegal information,thus ensuring national security and social stability.In this dissertation,iLBC(Internet Low Bit Rate Codec)speech streams oriented information steganography and steganalysis is studied,which mainly includes three aspects as following:(1)The existing research of iLBC coded speech streams oriented information steganography and steganalysis is investigated and analyzed.(2)Aiming at how to balance the embedded capacity,imperceptibility and undetectability,an iLBC speech steganography method based on gain quantization process is proposed.Firstly,the quantization table in the iLBC gain quantization process is reasonably divided,and then the corresponding groups are selected according to different secret information bits to quantize the gain.Finally,the gain quantization index is packaged and sent to the decoding end,and the opposite embedding process is performed from the decoding end.Secret information can be extracted.(3)For the gain quantization steganography,a steganalysis method based on Long Short-Term Memory(LSTM)is proposed.Firstly,it analyzes the correlation between the gain quantization index sequence in the iLBC speech streams and what changes after embedding the secret information,and then extracts the four most relevant correlation features,and selects one,two and four features respectively as the input of the LSTM network.Finally compare with the detection performance of different input for gain quantization steganography.