AttrLostGAN: Attribute Controlled Image Synthesis from Reconfigurable Layout and Style

TL;DR

AttrLostGAN enables fine-grained, attribute-specific control over individual objects in layout-based image synthesis, enhancing realism and diversity in complex scene generation.

Contribution

It extends layout-to-image models to incorporate attribute control for individual objects, improving detail customization without affecting the entire scene.

Findings

Successfully controls object appearance details

Works on synthetic and real-world datasets

Produces realistic, complex scene images

Abstract

Conditional image synthesis from layout has recently attracted much interest. Previous approaches condition the generator on object locations as well as class labels but lack fine-grained control over the diverse appearance aspects of individual objects. Gaining control over the image generation process is fundamental to build practical applications with a user-friendly interface. In this paper, we propose a method for attribute controlled image synthesis from layout which allows to specify the appearance of individual objects without affecting the rest of the image. We extend a state-of-the-art approach for layout-to-image generation to additionally condition individual objects on attributes. We create and experiment on a synthetic, as well as the challenging Visual Genome dataset. Our qualitative and quantitative results show that our method can successfully control the fine-grained details of individual objects when modelling complex scenes with multiple objects. Source code, dataset and pre-trained models are publicly available (https://github.com/stanifrolov/AttrLostGAN).