Detection Enhancement for Various Deepfake Types Based on Residual Noise and Manipulation Traces.
Jihyeon KangSang-Keun JiSangyeong LeeDaehee JangJong-Uk HouPublished in: IEEE Access (2022)
Keyphrases
- low signal to noise ratio
- detection method
- noisy environments
- detection algorithm
- signal detection
- automatic detection
- signal to noise ratio
- detection rate
- change detection
- noise model
- random noise
- false positives
- missing data
- image enhancement
- digital images
- detection scheme
- noisy data
- false detections
- residual error
- anti noise
- varying degrees
- multiple types
- real time
- speech recognition
- anomaly detection
- object detection
- image processing