Protecting the Copyright of Intelligent Transportation Systems Based on Zernike Moments

Abstract

Intelligent transportation systems are at risk of data misuse. Watermarking data that needs to be opened and shared can mitigate this problem, i.e., embedding signals into generated images, which are imperceptible to humans. Regardless of the shape of the image or any potential attacks it may undergo, the watermark can be detected by algorithms when needed. However, many watermarking schemes fail to resist attacks like cropping and translation, limiting their applicability in the intelligent transportation domain. To tackle these issues, we propose a dual watermarking framework based on Zernike moments for intelligent transportation systems, where a robust watermark and a periodic watermark are embedded in different planes of the cover image. Specifically, we propose a single-circle model (SCM) where Zernike moments are locally computed based on a circle centered at the image center with a radius proportional to image size for embedding the robust watermark. Since SCM is determined by measuring its center and radius, SCM is applicable to images of various sizes and shapes. Then we employ a robust combination of discrete wavelet transform (DWT) and discrete cosine transform (DCT) watermarking algorithms to embed the periodic watermark. For watermark extraction, we propose an efficient adaptive correction mechanism (AC) to recognize attack types and automatically relocate the embedding position of the watermark. By combining the above strategies, our proposed scheme can adaptively resist various attacks (e.g., random cropping, translation), which addresses the shortcomings of most existing watermarking schemes. We test the watermark using over 300 images of different sizes and shapes, and the experimental results prove that our proposed scheme achieves stronger robustness against various distortions with better invisibility, outperforming the state-of-the-art (SOTA) methods.