High-performance combination method of electric network frequency and phase for audio forgery detection in battery-powered devices
| Autor: | Ainuddin Wahid Abdul Wahab, Maryam Savari, Nor Badrul Anuar |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Battery (electricity)
021110 strategic defence & security studies business.product_category Computer science business.industry Forgery detection Speech recognition 0211 other engineering and technologies Phase (waves) 020206 networking & telecommunications 02 engineering and technology Pathology and Forensic Medicine Mobile phone Feature (computer vision) Laptop 0202 electrical engineering electronic engineering information engineering Computer vision Electric network Artificial intelligence business Combination method Law |
| Zdroj: | Forensic Science International. 266:427-439 |
| ISSN: | 0379-0738 |
| Popis: | Audio forgery is any act of tampering, illegal copy and fake quality in the audio in a criminal way. In the last decade, there has been increasing attention to the audio forgery detection due to a significant increase in the number of forge in different type of audio. There are a number of methods for forgery detection, which electric network frequency (ENF) is one of the powerful methods in this area for forgery detection in terms of accuracy. In spite of suitable accuracy of ENF in a majority of plug-in powered devices, the weak accuracy of ENF in audio forgery detection for battery-powered devices, especially in laptop and mobile phone, can be consider as one of the main obstacles of the ENF. To solve the ENF problem in terms of accuracy in battery-powered devices, a combination method of ENF and phase feature is proposed. From experiment conducted, ENF alone give 50% and 60% accuracy for forgery detection in mobile phone and laptop respectively, while the proposed method shows 88% and 92% accuracy respectively, for forgery detection in battery-powered devices. The results lead to higher accuracy for forgery detection with the combination of ENF and phase feature. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect