1 paper
Siding Zeng, Jiangyan Yi, Jianhua Tao +4
When the task of locating manipulation regions in partially-fake audio (PFA) involves cross-domain datasets, the performance of deep learning models drops significantly due to the…