Human behaviour-based automatic depression analysis using hand-crafted statistics and deep learned spectral features

Keyphrases

• Human Behavior 100%
• Spectral Features 100%
• Behavior-based 100%
• Sequence-level 100%
• Frame-level 100%
• Severity Estimation 100%
• Depression Analysis 100%
• Convolutional Neural Network 50%
• Audio 50%
• Diagnostic Process 50%
• Time Series Data 50%
• Detection Accuracy 50%
• Severe Mental Illness 50%
• Clinical Diagnosis 50%
• Computational Load 50%
• Redundant Information 50%
• Binary Classification 50%
• Hand Use 50%
• Depression Detection 50%
• Network Problems 50%
• Level Statistics 50%
• Sequence Descriptors 50%
• Traditional Clinical 50%
• Multichannel Behaviour 50%
• Multi-channel Time Series 50%
• Facial Action Units 50%
• Segment-level 50%
• Behavioural Correlation 50%
• Gaze Direction 50%
• Spectral Representation 50%
• Depression Severity 50%
• Regression Experiment 50%

Computer Science

• Convolutional Neural Network 100%
• Decision-Making 100%
• Final Decision 100%
• Analysis System 100%
• Binary Classification 100%
• Detection Accuracy 100%
• Computational Load 100%
• Spectral Representation 100%
• Spectral Feature 100%
• Gaze Direction 100%
• Time Series Data 100%
• facial action 100%

Engineering

• Multichannel 100%
• Spectral Feature 100%
• Convolutional Neural Network 50%
• Systems Analysis 50%
• Longer Term 50%
• Conducted Experiment 50%
• Redundant Information 50%
• Computational Load 50%
• Data Series 50%
• Gaze Direction 50%