S2IGAN: Speech-to-Image Generation via Adversarial Learning
This addresses the challenge of enabling unwritten languages to benefit from image generation technologies, though it appears incremental as it adapts existing adversarial learning methods to a new modality.
The paper tackles the problem of generating images from speech descriptions for unwritten languages, proposing S2IGAN, which synthesizes high-quality and semantically-consistent images from speech signals without using text, as demonstrated on CUB and Oxford-102 datasets.
An estimated half of the world's languages do not have a written form, making it impossible for these languages to benefit from any existing text-based technologies. In this paper, a speech-to-image generation (S2IG) framework is proposed which translates speech descriptions to photo-realistic images without using any text information, thus allowing unwritten languages to potentially benefit from this technology. The proposed S2IG framework, named S2IGAN, consists of a speech embedding network (SEN) and a relation-supervised densely-stacked generative model (RDG). SEN learns the speech embedding with the supervision of the corresponding visual information. Conditioned on the speech embedding produced by SEN, the proposed RDG synthesizes images that are semantically consistent with the corresponding speech descriptions. Extensive experiments on two public benchmark datasets CUB and Oxford-102 demonstrate the effectiveness of the proposed S2IGAN on synthesizing high-quality and semantically-consistent images from the speech signal, yielding a good performance and a solid baseline for the S2IG task.