1 个仓库
Applying time and frequency masking to audio spectrograms to increase robustness in speech recognition models.
Distinct from Spectrogram Generation: Distinct from Spectrogram Generation: focuses on the masking operation for augmentation, not the visualization process.
Explore 1 awesome GitHub repository matching data & databases · Spectrogram Masking. Refine with filters or upvote what's useful.
nlpaug is a data augmentation library designed to generate synthetic text, audio, and spectrogram data to improve the robustness of machine learning models. It functions as a textual data synthesizer and an audio signal augmentor, providing specialized tools to expand datasets through various transformation methods. The project distinguishes itself through its ability to orchestrate complex workflows using a pipeline orchestrator, which allows multiple augmentation functions to be chained together sequentially or randomly. It supports sophisticated text synthesis via back-translation, context
Transforms audio spectrograms using time and frequency masking to improve speech recognition robustness.