StackGAN: Text to Photo-Realistic Image Synthesis with Stacked Generative Adversarial Networks

TLDR

Synthesizing high‑quality images from text is a challenging computer‑vision task, and existing methods can roughly reflect descriptions but lack detailed, vivid object parts. This work introduces StackGAN, a framework that generates 256×256 photo‑realistic images conditioned on text descriptions. StackGAN decomposes the task into two stages: a Stage‑I GAN sketches primitive shapes and colors from text, while a Stage‑II GAN refines these sketches into high‑resolution images, and a Conditioning Augmentation technique is added to improve diversity and stabilize training. Experiments on benchmark datasets show that StackGAN significantly outperforms state‑of‑the‑art methods, rectifying Stage‑I defects and adding compelling details to produce photo‑realistic images.

Abstract

Synthesizing high-quality images from text descriptions is a challenging problem in computer vision and has many practical applications. Samples generated by existing textto- image approaches can roughly reflect the meaning of the given descriptions, but they fail to contain necessary details and vivid object parts. In this paper, we propose Stacked Generative Adversarial Networks (StackGAN) to generate 256.256 photo-realistic images conditioned on text descriptions. We decompose the hard problem into more manageable sub-problems through a sketch-refinement process. The Stage-I GAN sketches the primitive shape and colors of the object based on the given text description, yielding Stage-I low-resolution images. The Stage-II GAN takes Stage-I results and text descriptions as inputs, and generates high-resolution images with photo-realistic details. It is able to rectify defects in Stage-I results and add compelling details with the refinement process. To improve the diversity of the synthesized images and stabilize the training of the conditional-GAN, we introduce a novel Conditioning Augmentation technique that encourages smoothness in the latent conditioning manifold. Extensive experiments and comparisons with state-of-the-arts on benchmark datasets demonstrate that the proposed method achieves significant improvements on generating photo-realistic images conditioned on text descriptions.