Research On Digital Fingerprint Generation On Compressed Media Based On Hash Learning

In recent years,with the rapid development of computer network and digital imaging equipment,image data has experienced an exponential growth.It is of great significance to explore how to better store and utilize large-scale image datasets.In the era of big data,image storage,fast retrieval and content security are hot topics.Among them,hash learning and digital fingerprinting are important technical means of information retrieval and information security respectively.In the retrieval and application of large-scale datasets,there is little exploration and research on information security.This paper at first time combine the hash,data compression,and digital fingerprint in the storage and research of large-scale image datasets and propose a new method of digital fingerprint generation on compressed media based on hash learning.This method can not only improve the retrieval speed on large scale image datasets,but also reduce the data storage capacity,which is beneficial to data processing and transmission,and enhance the security of information.The main works are as follows:(1)We explore the image retrieval method based on hash learning,image compression and digital fingerprint mechanism.It is the first time to combine hash learning,data compression and digital fingerprint.In this article,we proposed a digital fingerprint generation method in the compression of the media based on hash learning.It will automatically compress the image data and embed the digital fingerprint in the process of the retrieval in large-scale image datasets.(2)The orthogonal fingerprint and error-correcting coding fingerprint are analyzed and studied and their complexity is compared in this work.In order to be better applied to big datasets and multi-user scenarios,we combine Reed Solomon code and hash code to construct a new fingerprint module.On the one hand,it improves the ability to resist the average collusion attack.On the other hand,by extracting the fingerprint information in the image,the manager can quickly separate the hash code in the digital fingerprint,and retrieve the original image without compression or other processing in the dataset with constant time complexity.