I see what you hear: a vision-inspired method to localize words
This provides a lightweight solution for keyword spotting in audio streams, though it is incremental as it applies existing methods to a new domain.
The paper tackles word localization in speech data by adapting visual object detection techniques, achieving a 94% reduction in model size and a 6.5% improvement in F1 score on the LibriSpeech dataset for 1000 words.
This paper explores the possibility of using visual object detection techniques for word localization in speech data. Object detection has been thoroughly studied in the contemporary literature for visual data. Noting that an audio can be interpreted as a 1-dimensional image, object localization techniques can be fundamentally useful for word localization. Building upon this idea, we propose a lightweight solution for word detection and localization. We use bounding box regression for word localization, which enables our model to detect the occurrence, offset, and duration of keywords in a given audio stream. We experiment with LibriSpeech and train a model to localize 1000 words. Compared to existing work, our method reduces model size by 94%, and improves the F1 score by 6.5\%.