Papers › ToLo: A Two-Stage, Training-Free Layout-To-Image Generation Framework For High-Overlap Layouts

ToLo: A Two-Stage, Training-Free Layout-To-Image Generation Framework For High-Overlap Layouts

3 Mar 2025arXiv:2503.01667archive 2025-07-28

Linhao Huang, Jing Yu

Recent training-free layout-to-image diffusion models have demonstrated remarkable performance in generating high-quality images with controllable layouts. These models follow a one-stage framework: Encouraging the model to focus the attention map of each concept on its corresponding region by defining attention map-based losses. However, these models still struggle to accurately follow layouts with significant overlap, often leading to issues like attribute leakage and missing entities. In this paper, we propose ToLo, a two-stage, training-free layout-to-image generation framework for high-overlap layouts. Our framework consists of two stages: the aggregation stage and the separation stage, each with its own loss function based on the attention map. To provide a more effective evaluation, we partition the HRS dataset based on the Intersection over Union (IoU) of the input layouts, creating a new dataset for layout-to-image generation with varying levels of overlap. Through extensive experiments on this dataset, we demonstrate that ToLo significantly enhances the performance of existing methods when dealing with high-overlap layouts. Our code and dataset are available here: https://github.com/misaka12435/ToLo.

PaperPDFCode

Code

misaka12435/tolo officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AttributeImage GenerationLayout-to-Image Generation

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AttentionDiffusionFocusSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections