Discriminative Sounding Objects Localization via Self-supervised Audiovisual Matching

Autor:	Hu, Di, Qian, Rui, Jiang, Minyue, Tan, Xiao, Wen, Shilei, Ding, Errui, Lin, Weiyao, Dou, Dejing
Rok vydání:	2020
Předmět:	Computer Science - Computer Vision and Pattern Recognition Computer Science - Machine Learning Computer Science - Multimedia Computer Science - Sound Electrical Engineering and Systems Science - Audio and Speech Processing
Druh dokumentu:	Working Paper
Popis:	Discriminatively localizing sounding objects in cocktail-party, i.e., mixed sound scenes, is commonplace for humans, but still challenging for machines. In this paper, we propose a two-stage learning framework to perform self-supervised class-aware sounding object localization. First, we propose to learn robust object representations by aggregating the candidate sound localization results in the single source scenes. Then, class-aware object localization maps are generated in the cocktail-party scenarios by referring the pre-learned object knowledge, and the sounding objects are accordingly selected by matching audio and visual object category distributions, where the audiovisual consistency is viewed as the self-supervised signal. Experimental results in both realistic and synthesized cocktail-party videos demonstrate that our model is superior in filtering out silent objects and pointing out the location of sounding objects of different classes. Code is available at https://github.com/DTaoo/Discriminative-Sounding-Objects-Localization. Comment: To appear in NeurIPS 2020. Previous Title: Learning to Discriminatively Localize Sounding Objects in a Cocktail-party Scenario
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2010.05466 Zobrazit plný text záznamu View this record from Arxiv