Toward Copyright Integrity and Verifiability via Multi-Bit Watermarking for Intelligent Transportation Systems

TL;DR

This work proposes ITSmark, a watermarking framework for intelligent transportation systems that leverages a copyright-derived multi-bit watermark space to verify data integrity and copyright, with a permission layer for secure extraction and traceability of tampered content. ITSmark embeds bits by partitioning a multi-bit space into segments assigned to next-token choices, allowing custom embedding positions and embedding ratios; extraction is reversible and requires the correct cipher data and private key. It supports permission verification, tamper location tracing, and unforgeability, achieving better data quality and reliability than baselines on multiple ITS and NLP datasets. The results indicate ITSmark provides practical copyright protection for ITS data while preserving data quality and enabling forensic analysis.

Abstract

Intelligent transportation systems (ITS) use advanced technologies such as artificial intelligence to significantly improve traffic flow management efficiency, and promote the intelligent development of the transportation industry. However, if the data in ITS is attacked, such as tampering or forgery, it will endanger public safety and cause social losses. Therefore, this paper proposes a watermarking that can verify the integrity of copyright in response to the needs of ITS, termed ITSmark. ITSmark focuses on functions such as extracting watermarks, verifying permission, and tracing tampered locations. The scheme uses the copyright information to build the multi-bit space and divides this space into multiple segments. These segments will be assigned to tokens. Thus, the next token is determined by its segment which contains the copyright. In this way, the obtained data contains the custom watermark. To ensure the authorization, key parameters are encrypted during copyright embedding to obtain cipher data. Only by possessing the correct cipher data and private key, can the user entirely extract the watermark. Experiments show that ITSmark surpasses baseline performances in data quality, extraction accuracy, and unforgeability. It also shows unique capabilities of permission verification and tampered location tracing, which ensures the security of extraction and the reliability of copyright verification. Furthermore, ITSmark can also customize the watermark embedding position and proportion according to user needs, making embedding more flexible.

Figures (11)

• Figure 1: The performance of some SOTA schemes and the proposed ITSmark. “$\uparrow$” and “$\downarrow$” represent the higher / lower the value, the better the result. The metrics are found in “Section \ref{['sec5']}”.
• Figure 2: The embedding and extraction processes of ITSmark. “Pubkey” and “Privkey” are the public key and the private key. They are generated by the encryption algorithm. The encryption algorithm can be arbitrary. When it is actually deployed, an encryption algorithm with efficient key management and distribution can be adopted, such as PKI that can effectively manage the distribution and verification of public keys or key lifecycle management tools to help automate key generation, distribution, update, revocation, and other operations. Before extraction, permission verification is necessary. Only after successful verification can the copyright be extracted.
• Figure 3: The working principle of the “Embedder” and “Extractor” in ITSmark. Here only a certain moment's embedding and extraction processes are shown. The context is input into the “Embedder” to obtain the next token containing the watermark information.
• Figure 4: Comparison of ITSmark (Ours) and baselines regarding Perplexity. The lower the Perplexity value, the better. The squares in the box represent the mean value, the horizontal line in the box represents the median value, and the gray circle represents the outlier.
• Figure 5: Comparison of ITSmark (Ours) and baselines regarding BERTScore and ROUGE. The higher the BERTScore and ROUGE values, the better.