Resmax: Detecting voice spoofing attacks with residual network and max feature map

Keyphrases

• Attack Detection System 50%
• Automatic Speech Recognition 100%
• Constant-Q Transform 50%
• Deep Learning Model 50%
• Deployment Constraints 50%
• Detection Accuracy 50%
• Ensemble Approach 50%
• Ensemble System 50%
• Equal Error Rate 100%
• Feature Map 100%
• Highly Accurate 50%
• Latency Constraint 50%
• Lightweight CNN 50%
• Lightweight System 50%
• Model Complexity 50%
• Real-world Deployment 100%
• Replay Attack Detection 50%
• Residual Network 100%
• Single Model 50%
• Skip Connection 50%
• Spoofing 100%
• Spoofing Attack Detection 50%
• U-Net 50%
• Voice Spoofing Attack 100%

Computer Science

• Convolutional Neural Network 50%
• Deep Learning Model 50%
• Detection Accuracy 50%
• Ensemble Approach 50%
• Equal Error Rate 100%
• Feature Map 100%
• Model Complexity 50%
• Residual Neural Network 100%
• Speaker Verification 100%
• Spoofing 100%
• Spoofing Attack 100%