Exploring Diversified Similarity with Kundaha

Autor(es):Mostrar menos -

Santos, Lucio F. D. ; Blanco, Gustavo ; de Oliveira, Daniel ; Traina, Agma J. M. ; Traina, Caetano, Jr. ; Bedo, Marcos V. N. ; Cuzzocrea, A ; Allan, J ; Paton, N ; Srivastava, D ; Agrawal, R ; Broder, A ; Zaki, M ; Candan, S ; Labrinidis, A ; Schuster, A ; Wang, H

• Tipo de documento: Artigo Científico
• Fonte: CIKM'18: PROCEEDINGS OF THE 27TH ACM INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT; v. N/A, p. 4-pg., 2018-01-01.

Resumo

Exploring large medical image sets by means of traditional similarity query criteria (e.g., neighborhood) can be fruitless if retrieved images are too similar among themselves. This demonstration introduces Kundaha, an exploration tool that assists experts in retrieving and navigating on results from a diversified similarity perspective of user-posed queries. Its implementation includes a wide set of metrics, descriptors, and indexes for enhancing query execution. Users can combine such features with diversified similarity criteria for the organized exploration of result sets and also employ relevance feedback cycles for finding new query-based viewpoints. (AU)

• Processo FAPESP: 16/17078-0 - Mineração, indexação e visualização de Big Data no contexto de sistemas de apoio à decisão clínica (MIVisBD)
• Beneficiário: Agma Juci Machado Traina
• Modalidade de apoio: Auxílio à Pesquisa - Temático