GreatSplicing: A Semantically Rich Splicing Dataset

TL;DR

GreatSplicing is a large-scale, high-quality dataset with diverse semantic categories designed to improve splicing forgery detection models and address overfitting and benchmarking issues in the field.

Contribution

The paper introduces GreatSplicing, a new dataset with 5,000 images across 335 semantic categories, enhancing the training and evaluation of splicing detection models.

Findings

Models trained on GreatSplicing show lower misidentification rates.

Detection models exhibit better cross-dataset generalization.

GreatSplicing provides a more comprehensive benchmark for future research.

Abstract

In existing splicing forgery datasets, the insufficient semantic variety of spliced regions causes trained detection models to overfit semantic features rather than learn genuine splicing traces. Meanwhile, the lack of a reasonable benchmark dataset has led to inconsistent experimental settings across existing detection methods. To address these issues, we propose GreatSplicing, a manually created, large-scale, high-quality splicing dataset. GreatSplicing comprises 5,000 spliced images and covers spliced regions across 335 distinct semantic categories, enabling detection models to learn splicing traces more effectively. Empirical results show that detection models trained on GreatSplicing achieve low misidentification rates and stronger cross-dataset generalization compared to existing datasets. GreatSplicing is now publicly available for research purposes at the following link.