Using spatial data mining to improve retrieval of animal sound recordings

• Grant number: 12/11395-3
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Effective date (Start): September 15, 2012
• Effective date (End): September 14, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Claudia Maria Bauzer Medeiros
• Grantee: Daniel Cintra Cugler
• Supervisor: Shashi Shekhar
• Host Institution: Instituto de Computação (IC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Research place: University of Minnesota (U of M), United States
• Associated to the scholarship: 11/19284-3 - Supporting the Collection and Curation of Biological Observation Databases, BP.DR

Abstract

For decades, biologists around the world have recorded animal sounds. As the number of records grows, so does the difficulty to manage them, presenting challenges to save, retrieve, share and manage sounds. These challenges are complicated by the fact that animal sound recordings have specific peculiarities, associated to the context in which the sound was recorded. For example, sounds emitted by individuals that are in groups may be different from ones emitted by isolated individuals. Though these characteristics may be relevant to biologists, they are seldom explicit in the recording metadata. This project presents our architecture for animal sound retrieval based on context analysis, considering factors such as environmental contexts, which are not treated by current systems. However, the architecture does not consider spatial features of the sound metadata. These features might hide interesting and useful knowledge to support the retrieval process. The goal of this project is to apply spatial data mining techniques to metadata associated to animal sound recordings. The idea is to discover hidden and interesting patterns between metadata and human made georeferenced elements (e.g., roads and train rails) in order to improve our architecture, by supporting biologists with information that can guide them on creating queries. (AU)