Complex Image Generation SwinTransformer Network for Audio Denoising
This addresses audio denoising for real-world applications, but it is incremental as it builds on existing image generation techniques.
The paper tackles audio denoising by converting it into an image generation task using a complex image generation SwinTransformer network, achieving better performance than state-of-the-art methods on two benchmark datasets.
Achieving high-performance audio denoising is still a challenging task in real-world applications. Existing time-frequency methods often ignore the quality of generated frequency domain images. This paper converts the audio denoising problem into an image generation task. We first develop a complex image generation SwinTransformer network to capture more information from the complex Fourier domain. We then impose structure similarity and detailed loss functions to generate high-quality images and develop an SDR loss to minimize the difference between denoised and clean audios. Extensive experiments on two benchmark datasets demonstrate that our proposed model is better than state-of-the-art methods.