Sound Prompted Semantic Segmentation

Sound prompted semantic segmentation aims to predict a segmentation mask given an audio prompt for the object in question. For example, given the sound of a car, the task is to segment the cars in the image.