Implementation of software prototype for P2P indexing and Collaborative Crawling

The focus of this deliverable is to describe the implementation of an integrated software for on-line indexing of audio-visual content into the SAPIR network. We illustrate several tools for pushing user-provided content into the SAPIR network. In particular we describe new SAPIR functionalities aimed at offering to the users the possibility to collaboratively push their audio-visual content into SAPIR enriched with user provided textual metadata. This integrated software is based on the prototypes implementations for P2P indexing, Caching, and Collaborative Crawling. We increased the size of the image collection used in most of the SAPIR experiments to 100 million images, and we validate the metric cache approach in presence of near duplicate queries