Learning Sight from Sound: Ambient Sound Provides Supervision for Visual Learning
This work addresses the challenge of unsupervised visual learning for computer vision researchers, though it is incremental as it extends a prior conference paper.
The paper tackles the problem of learning visual models without explicit labels by using ambient sound as a supervisory signal, resulting in a representation that performs comparably to state-of-the-art unsupervised methods on recognition tasks.
The sound of crashing waves, the roar of fast-moving cars -- sound conveys important information about the objects in our surroundings. In this work, we show that ambient sounds can be used as a supervisory signal for learning visual models. To demonstrate this, we train a convolutional neural network to predict a statistical summary of the sound associated with a video frame. We show that, through this process, the network learns a representation that conveys information about objects and scenes. We evaluate this representation on several recognition tasks, finding that its performance is comparable to that of other state-of-the-art unsupervised learning methods. Finally, we show through visualizations that the network learns units that are selective to objects that are often associated with characteristic sounds. This paper extends an earlier conference paper, Owens et al. 2016, with additional experiments and discussion.