Using Focused Crawlers with Obfuscation Techniques in the Audio Retrieval Domain

The detection of intellectual property violations in multimedia files poses a critical challenge for the Internet infrastructure, particularly in the context of very large document collections. The techniques employed to address these issues generally fall into two categories: proactive and reactive approaches. In this article we propose an approach that is both reactive and proactive, with the aim of preventing the deletion of legal uploads (or modifications of such files, such as remixes, parodies, and other edits) due to the presence of illegal uploads on a platform. We have developed a rule-based, obfuscating focused crawler that can work with files in the Audio Information Retrieval domain. Our model automatically scans multimedia files uploaded to a public collection only when a user submits a search query. We present experimental results obtained during tests on a well-known music collection, discussing the strength and efficiency of specific combinations of Neural Network-Similarity Scoring solutions.