Title:  Characterizing Acoustic Environments with Olaf and Elsa

Committee:

Dr. David Anderson, ECE, Chair , Advisor

Dr. Mark Davenport, ECE

Dr. Marilyn Wolf, ECE

Dr. Aaron Lanterman, ECE

Dr. Wayne Daley, GTRI

Dr. Polo Chau, CoC

Abstract:

The confluence of signal processing and machine learning has created many innovative technologies in popular research areas such as speech recognition.  However, many of the most successful methods are difficult to apply in areas that lack institutional support for research and creation of labeled data corpora.  The focus of this work is the development of signal processing and machine learning methods that can be practically implemented with less human effort, less need for large quantities of labeled data, and less computational cost.  Toward this goal, we have developed methods for outlier learning using augmented frozen dictionaries (OLAF) and estimating the likelihood of sparse approximations (ELSA) in the context of monitoring acoustic environments.  Both methods utilize sparse, dictionary-based representations to capture information about the structure of the data and have been tested for monitoring poultry production facilities.  They have proven effective in characterizing these environments and highlighting events or changes in the conditions present despite high levels of noise.  These tools have potential to help producers to better understand the effects different practices have on the animals and could lead to better animal well-being.