DocTamper
A deep learning model designed specifically for document tampering detection. Pre-trained on a large-scale document forgery dataset, it supports cross-dataset zero-shot transfer and is used to detect traces of manipulation in document images.
Timeline (last 90 days)
DocTamper is a deep learning model specifically designed for document tampering detection, pre-trained on large-scale document forgery datasets, with a certain degree of zero-shot cross-dataset transfer capability.
DocTamper's zero-shot transfer AUC on the Find It Again! dataset is approximately 0.50, indicating significant domain shift.
All Facts (2)
DocTamper is a deep learning model specifically designed for document tampering detection, pre-trained on large-scale document forgery datasets, with a certain degree of zero-shot cross-dataset transfer capability.
50%UnverifiedDocTamper's zero-shot transfer AUC on the Find It Again! dataset is approximately 0.50, indicating significant domain shift.
50%