Bimodal fusion emotion recognition method based on biological radar and voice information
A voice information and emotion recognition technology, applied in the field of emotion recognition, can solve the problems of low accuracy of emotional features, complex and changeable human emotions, etc., and achieve the effect of improving the degree of freedom of measurement
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0053] A dual-modal fusion emotion recognition method based on bio-radar and voice information, such as figure 1 described, including the following steps:
[0054] 1) Acquisition of speech signals and physiological signals: non-contact collection of natural speech and human body signals using microphones and radars;
[0055] Radar refers to frequency modulation continuous wave radar, which uses linear frequency modulation technology to transmit sawtooth waves;
[0056] The microphone refers to a digital MEMS microphone that outputs a 1 / 2 cycle pulse density modulated digital signal;
[0057]2) Signal preprocessing: Preprocess the signals of the two modalities of physiology and speech, including heartbeat signals, breathing signals, and speech signals, so that they meet the input requirements of the corresponding models of different modalities;
[0058] 3) Emotional feature extraction: perform feature extraction on the preprocessed heartbeat signal, breathing signal, and voic...
Embodiment 2
[0062] Physiological information processing flow, such as figure 2 described, including the following steps:
[0063] 1) Use frequency-modulated continuous wave radar to obtain human physiological sign signals, perform band-pass filtering on the original human physiological sign signals, filter out drift signals below 0.2 Hz and noise signals above 2 Hz in the original sign signals, and convert 0.2 Hz-0.9 Hz signals to It is classified as a respiratory signal, and the signal of 0.9Hz-2.0Hz is classified as a heartbeat signal;
[0064] 2) Extract the time feature, waveform feature and frequency domain feature of the respiratory signal; extract the time feature, waveform feature and frequency domain feature of the heartbeat signal;
[0065] 3) PCA dimensionality reduction is carried out on the extracted respiratory signal and heartbeat signal features, and reduced to two-dimensional data to obtain physiological features;
[0066] 4) Input the physiological feature data to be ...
Embodiment 3
[0068] Voice information processing flow, such as image 3 described, including the following steps:
[0069] 1) Utilize the digital MEMS microphone to obtain the human voice signal, and pre-emphasize the human voice signal through a digital filter, and output the pre-emphasized voice data;
[0070] 2) utilizing short-term analysis technology to carry out frame processing to the voice data after the pre-emphasis, and obtain the time series of voice feature parameters;
[0071] 3) Using the Hamming window function to perform windowing processing on the speech feature parameter time series to obtain speech windowing data
[0072] 4) utilize double-threshold comparison method to carry out endpoint detection to described voice window data, obtain the voice data after preprocessing;
[0073] 5) Carry out short-time Fourier transform to the speech data after the preprocessing, draw speech spectrogram;
[0074] 6) Extracting speech feature data to obtain speech emotion features; ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More