Move2Hear: Active Audio-Visual Source Separation
This addresses the challenge of separating overlapping audio sources in noisy environments for applications like augmented reality and mobile robotics, though it is incremental as it builds on existing audio-visual methods.
The paper tackles the problem of active audio-visual source separation, where an agent moves to isolate sounds from a target object, and demonstrates a reinforcement learning approach that finds minimal movement sequences for maximal separation payoff in simulated 3D environments.
We introduce the active audio-visual source separation problem, where an agent must move intelligently in order to better isolate the sounds coming from an object of interest in its environment. The agent hears multiple audio sources simultaneously (e.g., a person speaking down the hall in a noisy household) and it must use its eyes and ears to automatically separate out the sounds originating from a target object within a limited time budget. Towards this goal, we introduce a reinforcement learning approach that trains movement policies controlling the agent's camera and microphone placement over time, guided by the improvement in predicted audio separation quality. We demonstrate our approach in scenarios motivated by both augmented reality (system is already co-located with the target object) and mobile robotics (agent begins arbitrarily far from the target object). Using state-of-the-art realistic audio-visual simulations in 3D environments, we demonstrate our model's ability to find minimal movement sequences with maximal payoff for audio source separation. Project: http://vision.cs.utexas.edu/projects/move2hear.