Project # 45
Canine Age and Gender Classification based on their barks

Team Lead:	Hridayesh Lekhak
Other Team Members:
Faculty Advisor(s):	Dr. Kenny Zhu
Department:	CSE
Project Type:	REU
Title:	Canine Age and Gender Classification based on their barks
Abstract:	Understanding canine vocalizations provides valuable insights into their communication, emotional states and biological attributes. In this study, we introduce the Canine Vocalization Dataset, the largest curated collection of pure dog barks, accompanied by metadata such as breed, age and biological sex. This dataset facilitates research on the evolution of vocal patterns as dogs transition through different life stages and the acoustic and phonetic differences between male and female dogs across breeds. To establish benchmarks, we explore age classification using acoustic features such as MFCCs, GeMAPS, and HuBERT embeddings. Additionally, we analyze discrete phonetic units of dog barks to study the underlying structure of their vocal communication. By leveraging both acoustic and phonetic analyses, this research advances our understanding of canine vocalization patterns and their relation to age and gender.
Poster:	View Poster #45 (PDF)
Judges Scoring:	Submit Evaluation (judges only)

Project # 45Canine Age and Gender Classification based on their barks