There has been great interest in image forensics in recent years. However, most of the existing research focuses on detecting a certain tampering operation, which means that the introduced features usually depend on the investigated operation and only binary classification is considered. Given the case where the image tampering process involves diverse processing operations, we propose a decision fusion method for identifying tampering operations in operator chains in this work. The proposed method permits the integration of knowledge provided by available image forensic algorithms. Under this method, a similarity coefficient function is introduced to assign the weight of the output of each forensic classifier. Then, we utilize a combination rule based on local conflict management to merge these outputs. Comparison with the previous works shows an improvement in operations identification accuracy when an image has experienced multiple falsifications.

近年来，人们对图像取证有了极大的兴趣。然而，大多数现有研究集中于检测某种篡改操作，这意味着引入的功能通常取决于所研究的操作，并且仅考虑二进制分类。考虑到图像篡改过程涉及多种处理操作的情况，我们提出了一种决策融合方法，用于在这项工作中识别操作员链中的篡改操作。所提出的方法允许由可用的图像取证算法提供的知识的整合。在这种方法下，引入了一个相似系数函数来分配每个法证分类器的输出权重。然后，我们利用基于本地冲突管理的组合规则来合并这些输出。