基于多信息引导和渐进掩码Transformer的小篡改区域定位方法
By employing a small-tampering region localization method guided by multiple information and progressive masking Transformer, and by using edge and region information to guide the process, and refining and aggregating features layer by layer, the problem of unsatisfactory localization in small-region tampering is solved, achieving high-precision and robust localization results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HEBEI UNIV OF TECH
- Filing Date
- 2024-01-12
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods struggle to accurately locate small-area tampering in images, lack research on small-area tampering, and lack regional and edge guidance information, making it difficult to capture details of small-area tampering.
A small tampering region localization method based on multi-information guidance and progressive mask Transformer is adopted. Edge and region information are used to guide the localization of small tampering regions. Edge and region guidance information is generated through edge generation module and frequency domain guided region generation module, and features are refined and aggregated layer by layer using mask Transformer.
It achieves precise localization of small tampered regions, improves localization accuracy and robustness, outperforms existing methods on multiple tampered datasets, and has good recognition robustness.
Smart Images
Figure CN117876704B_ABST