Sound Classification and Localization Based on Biology Hearing Models and Multiscale Vector Quantization
Abstract
Acoustic Vehicle Classification Objectives and Challenges: 1. Develop systematic methodologies and algorithms; not ad hoc 2. Robust Target ID (with respect to environment, terrain, speed) 3. Algorithms for combined DOA (localization) and target ID 4. Localization assisted ID 5. ID assisted localization 6. Multi-target detection, ID and DOA separation of closely spaced targets 7. Robust feature extraction from auditory models; dynamic DOA and ID 8. Algorithm evaluation in the field and comparison against conventional algorithms for detection, DOA and ID.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 24, 1999
- Accession Number
- ADA512921
Entities
People
- John Baras
Organizations
- University of Maryland