A noise monitoring digital audio tamper detection method based on power grid frequency signal
By using a lightweight U-shaped visual Transformer architecture and a deep learning model with a local-global-local attention mechanism, the problems of low automation and low accuracy in digital audio tampering detection are solved, achieving efficient and accurate tampering detection under low signal-to-noise ratio conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUNAN UNIV
- Filing Date
- 2026-05-07
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies for detecting digital audio tampering suffer from several drawbacks. They rely on manual parameter adjustments and feature engineering, resulting in low automation, low accuracy in locating tampering boundaries, difficulty in achieving precise point-by-point annotation, insufficient robustness, limited detection capabilities under low signal-to-noise ratio conditions, and low computational efficiency, making it difficult to meet the real-time requirements of large-scale data review.
By employing a lightweight U-shaped vision Transformer architecture (Mobile U-ViT 1D) combined with a local-global-local attention mechanism, and using a deep learning model to automatically learn tampering feature patterns, the tampering detection of power grid frequency signals is transformed into a one-dimensional time series segmentation task, achieving precise point-by-point annotation.
Significant improvements have been achieved in detection accuracy, automation, computational efficiency, and robustness. It can adapt to the detection of complex tampering types under low signal-to-noise ratio conditions and meet the real-time requirements of large-scale data auditing.
Smart Images

Figure CN122173839A_ABST