Yuseung Lee

2 accepted papers

2024

GrounDiT: Grounding Diffusion Transformers via Noisy Patch Transplantation

NeurIPS 2024poster

We introduce GrounDiT, a novel training-free spatial grounding technique for text-to-image generation using Diffusion Transformers (DiT). Spatial grounding with bounding boxes has gained attention for its simplicity and versatility, allowing for enhanced user control in image generation. However, pr…

Cited by 2 Source ↗PDF ↗code ↗Scholar ↗

2023

SyncDiffusion: Coherent Montage via Synchronized Joint Diffusions

Yuseung Lee, Kunho Kim, Hyunjin Kim, Minhyuk Sung

NeurIPS 2023poster

The remarkable capabilities of pretrained image diffusion models have been utilized not only for generating fixed-size images but also for creating panoramas. However, naive stitching of multiple images often results in visible seams. Recent techniques have attempted to address this issue by perform…

Cited by 55 Source ↗PDF ↗code ↗Scholar ↗