arXiv subjects

Haoyu Lu

Publications and source records attributed to Haoyu Lu.

38 records · Page 3Linked to original sources

arXiv2022

Image Fragile Watermarking Algorithm Based on Deneighborhood Mapping

To address the security risk caused by fixed offset mapping and the limited recoverability of random mapping used in image watermarking, we propose an image self-embedding fragile watermarking algorithm based on deneighborhood mapping. First, the image is divided into several 2*2 blocks, and authentication watermark and recovery watermark are generated based on the average value of the image blocks. Then, the denighborhood mapping is implemented as, for each image block, its mapping block is randomly selected outside it's neighborhood whose size is specified by a parameter. Finally, the authentication watermark and the recovery watermark are embedded in the image block itself and its corresponding mapping block. Theoretical analysis indicates that in the case of continuous region tampering, the proposed watermarking method can achieve better the recovery rate of the tampered image block than the method based on the random mapping. The experimental results verify the rationality and effectiveness of the theoretical analysis. Moreover, compared with the existing embedding algorithms based on random mapping, chaos mapping and Arnold mapping, in the case of continuous region tampering, the average recovery rate of the tampered region achieved by the proposed algorithm is higher.

cs.CR↗

arXiv2021

WenLan: Bridging Vision and Language by Large-Scale Multi-Modal Pre-Training

Multi-modal pre-training models have been intensively explored to bridge vision and language in recent years. However, most of them explicitly model the cross-modal interaction between image-text pairs, by assuming that there exists strong semantic correlation between the text and image modalities. Since this strong assumption is often invalid in real-world scenarios, we choose to implicitly model the cross-modal correlation for large-scale multi-modal pre-training, which is the focus of the Chinese project `WenLan' led by our team. Specifically, with the weak correlation assumption over image-text pairs, we propose a two-tower pre-training model called BriVL within the cross-modal contrastive learning framework. Unlike OpenAI CLIP that adopts a simple contrastive learning method, we devise a more advanced algorithm by adapting the latest method MoCo into the cross-modal scenario. By building a large queue-based dictionary, our BriVL can incorporate more negative samples in limited GPU resources. We further construct a large Chinese multi-source image-text dataset called RUC-CAS-WenLan for pre-training our BriVL model. Extensive experiments demonstrate that the pre-trained BriVL model outperforms both UNITER and OpenAI CLIP on various downstream tasks.

cs.CV↗