AIbase
EN
Home
AST ASVspoof5 Synthetic Voice Detection
A

AST ASVspoof5 Synthetic Voice Detection

Developed by MattyB95
A synthetic speech detection model fine-tuned based on MIT/ast-finetuned-audioset-10-10-0.4593, used to identify whether an audio is synthetic speech.
Audio Classification
Transformers
Open Source License:Bsd-3-clause#Voice Anti-spoofing#High-precision Detection#Audio Classification
Downloads 281
Release Time : 7/20/2024

Model Overview

This model is an audio classification model specifically designed for detecting synthetic speech, demonstrating high accuracy and F1 score on the validation set.

Model Features

High Accuracy
Achieves 83.33% accuracy on the validation set, demonstrating excellent performance.
Balanced Performance
F1 score reaches 0.8892, indicating a good balance between precision and recall.
Fine-tuning Optimization
Fine-tuned based on a pre-trained model to adapt to the synthetic speech detection task.

Model Capabilities

Audio Classification
Synthetic Speech Detection

Use Cases

Security Verification
Voice Authentication System
Used to detect potential synthetic speech attacks in voice authentication systems.
Effectively identifies synthetic speech, reducing the risk of system spoofing.
Content Moderation
Fake Audio Detection
Identifies synthetic speech content on social media or news platforms.
Helps platforms filter out potentially fake or misleading audio content.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase