# ObjBlur: A Curriculum Learning Approach With Progressive Object-Level Blurring for Improved Layout-to-Image Generation

arXiv: https://arxiv.org/abs/2404.07564 · HTML: https://mlsanity.com/paper/2404.07564/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 38 references that are in the map.
- First posted on arXiv 11 Apr 2024.

## Builds on (most cited first, 27 of 27)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Progressive Growing of GANs for Improved Quality, Stability, and Variation](https://mlsanity.com/paper/1710.10196/) (2017)
- [Improved Techniques for Training GANs](https://mlsanity.com/paper/1606.03498/) (2016)
- [Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations](https://mlsanity.com/paper/1602.07332/) (2016)
- [Training Generative Adversarial Networks with Limited Data](https://mlsanity.com/paper/2006.06676/) (2020)
- [COCO-Stuff: Thing and Stuff Classes in Context](https://mlsanity.com/paper/1612.03716/) (2016)
- [Tackling the Generative Learning Trilemma with Denoising Diffusion GANs](https://mlsanity.com/paper/2112.07804/) (2021)
- [LayoutDiffusion: Controllable Diffusion Model for Layout-to-image Generation](https://mlsanity.com/paper/2303.17189/) (2023)
- [Diffusion-GAN: Training GANs with Diffusion](https://mlsanity.com/paper/2206.02262/) (2022)
- [Generative Modelling With Inverse Heat Dissipation](https://mlsanity.com/paper/2206.13397/) (2022)
- [Classification Accuracy Score for Conditional Generative Models](https://mlsanity.com/paper/1905.10887/) (2019)
- [Consistency Regularization for Generative Adversarial Networks](https://mlsanity.com/paper/1910.12027/) (2019)
- [Image Generation from Layout](https://mlsanity.com/paper/1811.11389/) (2018)
- [A Comprehensive Survey of Image Augmentation Techniques for Deep Learning](https://mlsanity.com/paper/2205.01491/) (2022)
- [Image Synthesis From Reconfigurable Layout and Style](https://mlsanity.com/paper/1908.07500/) (2019)
- [Image Augmentations for GAN Training](https://mlsanity.com/paper/2006.02595/) (2020)
- [Improved Consistency Regularization for GANs](https://mlsanity.com/paper/2002.04724/) (2020)
- [Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Analysis and a New Strategy](https://mlsanity.com/paper/2004.00448/) (2020)
- [Object-Centric Image Generation from Layouts](https://mlsanity.com/paper/2003.07449/) (2020)
- [Context-Aware Layout to Image Generation with Enhanced Object Appearance](https://mlsanity.com/paper/2103.11897/) (2021)
- [Progressive Deblurring of Diffusion Models for Coarse-to-Fine Image Synthesis](https://mlsanity.com/paper/2207.11192/) (2022)
- [Image Difficulty Curriculum for Generative Adversarial Networks (CuGAN)](https://mlsanity.com/paper/1910.08967/) (2019)
- [On-line Adaptative Curriculum Learning for GANs](https://mlsanity.com/paper/1808.00020/) (2018)
- [Language Generation with Recurrent Generative Adversarial Networks without Pre-training](https://mlsanity.com/paper/1706.01399/) (2017)
- [Adversarial Learning of Semantic Relevance in Text to Image Synthesis](https://mlsanity.com/paper/1812.05083/) (2018)
- [AttrLostGAN: Attribute Controlled Image Synthesis from Reconfigurable Layout and Style](https://mlsanity.com/paper/2103.13722/) (2021)
- [Improved Training with Curriculum GANs](https://mlsanity.com/paper/1807.09295/) (2018)

## Built on by (most cited first, 1 of 1)

- [CMA-OT: Hierarchical Expert Supervision for Dance-to-Music Generation](https://mlsanity.com/paper/2609.13118/) (2026)

## Abstract (arXiv)

We present ObjBlur, a novel curriculum learning approach to improve layout-to-image generation models, where the task is to produce realistic images from layouts composed of boxes and labels. Our method is based on progressive object-level blurring, which effectively stabilizes training and enhances the quality of generated images. This curriculum learning strategy systematically applies varying degrees of blurring to individual objects or the background during training, starting from strong blurring to progressively cleaner images. Our findings reveal that this approach yields significant performance improvements, stabilized training, smoother convergence, and reduced variance between multiple runs. Moreover, our technique demonstrates its versatility by being compatible with generative adversarial networks and diffusion models, underlining its applicability across various generative modeling paradigms. With ObjBlur, we reach new state-of-the-art results on the complex COCO and Visual Genome datasets.