Generating Realistic Images from In-the-wild Sounds

Representing wild sounds as images is an important but challenging task due to the lack of paired datasets between sound and images and the significant differences in the characteristics of these two modalities. Previous studies have focused on generating images from sound in limited categories or music. In this paper, we propose a novel approach to generate images from in-the-wild sounds. First, we convert sound into text using audio captioning. Second, we propose audio attention and sentence attention to represent the rich characteristics of sound and visualize the sound. Lastly, we propose a direct sound optimization with CLIPscore and AudioCLIP and generate images with a diffusion-based model. In experiments, it shows that our model is able to generate high quality images from wild sounds and outperforms baselines in both quantitative and qualitative evaluations on wild audio datasets.

Overall architecture of our model. Our model consists of two stages. Stage (a) is the step of initializing the W vector with attentions, and Stage (b) is direct sound optimization and image generation process. Process of direct sound optimization. When initializing latent z with attentions, we optimize in the blue area, whereas without using attention, we optimize in the yellow area. To solve the problem of local minimum and to represent the rich features of sound, attentions are necessary.