Seminář: Similarity Search in Non-Text Data

Due to extensive digitalization and use of computers connected in networks, search in many non-text data collections is based on specific similarity metrics. The Multi Feature Indexing Network (MUFIN) represents a scalable and extensible similarity search platform for many applications. Its extensibility is achieved by accepting the metric space model of similarity – the technology works for any metric distance measure and can serve applications as diverse as biology, security, geography, multimedia, data cleaning and integration, etc. In order to scale into billions of objects searched on-line for hundreds of queries, structured peer-to-peer (P2P) similarity search networks are applied. In order to tune performance, MUFIN keeps a clear separation between the logical P2P structure and the hardware physical infrastructure which is used as service. MUFIN’s capability will be illustrated on a 100 million image collection searched by an aggregation of five different MPEG-7 descriptors.