Automatic depression recognition by intelligent speech signal processing:A systematic survey

Depression has become one of the most common mental illnesses in the world.For better prediction and diagnosis,methods of automatic depression recognition based on speech signal are constantly proposed and updated,with a transition from the early traditional methods based on hand‐crafted features to the application of architectures of deep learning.This paper systematically and precisely outlines the most prominent and up‐to‐date research of automatic depression recognition by intelligent speech signal processing so far.Furthermore,methods for acoustic feature extraction,algorithms for classification and regression,as well as end to end deep models are investigated and analysed.Finally,general trends are summarised and key unresolved issues are identified to be considered in future studies of automatic speech depression recognition.

Underwater Noise Target Recognition Based on Sparse Adversarial Co-Training Model with Vertical Line Array

The automatic identification of underwater noncooperative targets without label records remains an arduous task considering the marine noise interference and the shortage of labeled samples.In particular,the data-driven mechanism of deep learning cannot identify false samples,aggravating the difficulty in noncooperative underwater target recognition.A semi-supervised ensemble framework based on vertical line array fusion and the sparse adversarial co-training algorithm is proposed to identify noncooperative targets effectively.The sound field cross-correlation compression(SCC)feature is developed to reduce noise and computational redundancy.Starting from an incomplete dataset,a joint adversarial autoencoder is constructed to extract the sparse features with source depth sensitivity,aiming to discover the unknown underwater targets.The adversarial prediction label is converted to initialize the joint co-forest,whose evaluation function is optimized by introducing adaptive confidence.The experiments prove the strong denoising performance,low mean square error,and high separability of SCC features.Compared with several state-of-the-art approaches,the numerical results illustrate the superiorities of the proposed method due to feature compression,secondary recognition,and decision fusion.

Signal processing in underwater acoustic communication system for manned deep submersible "Jiaolong"

For the first time in the world, underwater acoustic transmission of images, human voice, data and texts between vehicle under 7000 m depth and surface ship was accomplished by underwater acoustic communication system of manned deep submersible Jiaolong＇. In this paper, signal processing in underwater acoustic communication system for manned deep submersible ＂Jiaolong＂ is introduced. （1） Four communication methods are integrated to meet different needs： 1） coherent underwater acoustic communication, with a variable transmission rate from 5 kbps to 15 kbps, to transmit images. 2） Non-coherent underwater acoustic com- munication, with a transmission rate 300 bps, to transmit texts, instructions, and sensor data. 3） Spread spectrum underwater acoustic communication, with a transmission rate 16 bps, to transmit instructions. 4） Underwater voice communication, using single sideband modulation to transmit hmnan voice. （2） Signal processing method in coherent communication mainly consists of concatenation of decision feedback equalizer and Turbo decoder, and wavelet based image compression with fixed length coding. In the equalizer, Doppler compensation, multi- channel combining and equalizer coefficients updating are all using fast self-optimized adaptive algorithm. （3） A linear hydrophone array is lowered from the mother ship to certain depth, and spatial diversity combining technology is adopted. （4） Diving trials of ＂Jiaolong＂ were carried out in Pacific Ocean. The communication range can cover nearly all ocean depth. One optical/acoustic image can be transmitted in 7 or 14 seconds.

Signal processing for noncoherent underwater acoustic communication approaching channel capacity

Noncoherent underwater acoustic communication channel in adverse conditions is modeled as a phase-random Rayleigh fading channel,and its capacity curve is derived.To approach the channel capacity curve,the concatenated code of the nonbinary LDPC code and the constant weight code is proposed for noncoherent communication which can late be iteratively decoded in the probability domain.Without information of channel amplitude or phase in the receiver,statistic parameters of the respective signal and noise bins were estimated based on the moment estimation method,the posterior probabilities of the constant weight code words were further calculated,and the nonbinary LDPC code was decoded with the nonbinary factor graph algorithm.It is verified by simulations that by utilizing the proposed concatenated code and its processing algorithm,gap to channel capacity curve is reduced by 3 dB when compared to the existing method.Underwater communication experiments were carried out in both deep ocean（vertical communication,5 km）and shallow lake（horizontal communication,near 3 km,delay spread larger than 50 ms）,in which the signal frequency band was 6 kHz to10 kHz,and the data transmission rate Was 357 bps.The proposed scheme can work properly in both experiments with a signal-to-noise ratio threshold of 2 dB.The performance of the proposed algorithm Was well verified by the experiments.

The laboratory of acoustics,speech and signal processing at the institute of acoustics

The Laboratory of Acoustics,Speech and Signal Processing(LASSP),theunique and superior national key laboratory of ASSP in China,has been foundedat the Inst.of Acoustics,Academia Sinica,Beijing PRC.After three years ofefforts,the construction of the LASSP has been completed successfully and thecertain capability of performing frontier research projects in fundamental theory andapplied technology of sound field and acoustic signal processing has ben formed.A fiexible and complete experimental acoustic signal processing system hasbeen set up in the LASSP.With the remarkable advantage of real time signalprocessing and resource sharing,a wide range of research projects in the field ofASSP can be conducted in the laboratory.The Signal Processing Center of theLASSP is well equipped with many computer research facilities including the

Short Baseline Positioning with an Improved Time Reversal Technique in a Multi-path Channel

The existence of a multi-path channel under the water greatly decreases the accuracy of the short baseline positioning system.In this paper,the application of a time reversal mirror to the short baseline positioning system was investigated.The time reversal mirror technique allowed the acoustic signal to better focus in an unknown environment,which effectively reduced the expansion of multi-path acoustic signals as well as improved the signal focusing.The signal-to-noise ratio（SNR） of the time reversal operator greatly increased and could be obtained by ensonifying the water.The technique was less affected by the environment and therefore more applicable to a complex shallow water environment.Numerical simulations and pool experiments were used to demonstrate the efficiency of this technique.

Causal propagation of signals in strangeon matter

The state equation for strangeon matter is very stiff due to the non-relativistic nature of its particles and their repulsive interaction, such that pulsar masses as high as ～ 3M would be expected. However, an adiabatic sound speed, cs = √P/ρ, is usually superluminal in strangeon matter, and the dynamic response of a strangeon star （e.g., binary merger） is not tractable in numerical simulations. In this study, we examined signal propagation in strangeon matter and calculate the actual propagation speed, Csignal. We found that the causality condition, Csignal 〈 c, is satisfied and the signal speed is presented as a function of stellar radius.

Automatic Artist Recognition of Songs for Advanced Retrieval

Automatic recognition of artists is very important in acoustic music indexing, browsing, and contentbased acoustic music retrieving, but synchronously it is still a challenging errand to extract the most representative and salient attributes to depict diversiform artists. In this paper, we developed a novel system to complete the reorganization of artist automatically. The proposed system can efficiently identify the artist's voice of a raw song by analyzing substantive features extracted from both pure music and singing song mixed with accompanying music. The experiments on different genres of songs illustrate that the proposed system is possible.