Learning to Generate Images of Outdoor Scenes from Attributes and Semantic Layouts
This work addresses the challenge of automatic image synthesis for outdoor scenes, which is incremental as it builds on existing conditioning methods but applies them to a new domain.
The paper tackles the problem of generating realistic outdoor scene images by using a deep conditional generative adversarial network that integrates semantic layouts and scene attributes as conditioning variables, achieving clear object boundaries and variations like day-night and sunny-foggy conditions.
Automatic image synthesis research has been rapidly growing with deep networks getting more and more expressive. In the last couple of years, we have observed images of digits, indoor scenes, birds, chairs, etc. being automatically generated. The expressive power of image generators have also been enhanced by introducing several forms of conditioning variables such as object names, sentences, bounding box and key-point locations. In this work, we propose a novel deep conditional generative adversarial network architecture that takes its strength from the semantic layout and scene attributes integrated as conditioning variables. We show that our architecture is able to generate realistic outdoor scene images under different conditions, e.g. day-night, sunny-foggy, with clear object boundaries.