GreatSplicing is an extensive splicing forgery dataset. Compared with previous datasets in the field of splicing detection, GreatSplicing features rich varieties of semantics in spliced regions. In the production process, the amount of each semantic type was strictly controlled to ensure that the same semantic did not appear over and over. Additionally, every spliced image in GreatSplicing was manually created to ensure high authenticity, which was in line with practical splicing manipulations by image tampering software such as Adobe Photoshop. GreatSplicing comprises 5,000 spliced images, and more details can be found in the paper GreatSplicing: A Semantically Rich Splicing Dataset (VISITED: 0 times). In addition, we have also produced medical forgery dataset, which you can obtain through this site.