Joon Son Chung

FlexiAST allows models to adapt to various audio patch sizes efficiently.

Improving the way we identify sound sources using audio-visual data.

A new method improves speaker verification by managing session variability effectively.

This article discusses an automated method for generating movie trailers efficiently.

New methods improve video summarization using large datasets and advanced models.

ElasticAST allows processing of variable length audio efficiently without losing important details.

A study on improving sound source localization by better using audio and visual information.

An overview of advancements in speaker recognition through the VoxCeleb Challenge.