In image filtering methods,image features are analyzed and matched to identify the potential tampering sources and then to analyze the derivative relationship between the tampered images and their sources.In this paper,we propose a novel image filtering method based on object detection constraints by integrating the object detection(YOLO v5)and the distributed SURF indexing,where the object detection is adopted to constrain the feature retrieval scope and the distributed local feature extraction is utilized to improve the efficiency and accuracy of image feature retrieval.In addition,a scalable con-strained image splicing detection and localization network is employed for deep matching on the basis of image feature retrieval,allowing for precise identification of tampered areas and their sources.Experi-ment results demonstrate that the proposed method achieves significant improvements in tampered image retrieval accuracy and query efficiency.
维普
万方数据