Iterative Outlier Removal Method Using In-Cluster Variance Changes in Multi-Microphone Array Sound Source Localization.
○Daniel Gabriel,小島諒介,干場功太郎(東工大),中臺一博(東工大/ホンダRIJ)
In recent studies sound based localization plays a great role in fields such as Robot Audition and Spatial Bird Song Localization. It is difficult to distinguish between a valid sound source and environmental noises. This paper proposes an outlier-robust sound source localization method using multiple microphone arrays. Proposed algorithm has been tested on three datasets: a simulation dataset and two datasets of bird songs recorded in the real environment with different microphone array layouts. The algorithm has been compared with other outlier extraction techniques. We confirmed the effectiveness of the proposed algorithm.

footer 著作権について 倫理綱領 プライバシーポリシー セキュリティ 情報処理学会