Saved in:
Bibliographic Details
Main Authors: Qi, Fan, Duan, Yu, Xu, Changsheng
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.21069
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915215084355584
author Qi, Fan
Duan, Yu
Xu, Changsheng
author_facet Qi, Fan
Duan, Yu
Xu, Changsheng
contents Recent advances in text-guided diffusion models have revolutionized conditional image generation, yet they struggle to synthesize complex scenes with multiple objects due to imprecise spatial grounding and limited scalability. We address these challenges through two key modules: 1) Janus-Pro-driven Prompt Parsing, a prompt-layout parsing module that bridges text understanding and layout generation via a compact 1B-parameter architecture, and 2) MIGLoRA, a parameter-efficient plug-in integrating Low-Rank Adaptation (LoRA) into UNet (SD1.5) and DiT (SD3) backbones. MIGLoRA is capable of preserving the base model's parameters and ensuring plug-and-play adaptability, minimizing architectural intrusion while enabling efficient fine-tuning. To support a comprehensive evaluation, we create DescripBox and DescripBox-1024, benchmarks that span diverse scenes and resolutions. The proposed method achieves state-of-the-art performance on COCO and LVIS benchmarks while maintaining parameter efficiency, demonstrating superior layout fidelity and scalability for open-world synthesis.
format Preprint
id arxiv_https___arxiv_org_abs_2503_21069
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Efficient Multi-Instance Generation with Janus-Pro-Dirven Prompt Parsing
Qi, Fan
Duan, Yu
Xu, Changsheng
Computer Vision and Pattern Recognition
Recent advances in text-guided diffusion models have revolutionized conditional image generation, yet they struggle to synthesize complex scenes with multiple objects due to imprecise spatial grounding and limited scalability. We address these challenges through two key modules: 1) Janus-Pro-driven Prompt Parsing, a prompt-layout parsing module that bridges text understanding and layout generation via a compact 1B-parameter architecture, and 2) MIGLoRA, a parameter-efficient plug-in integrating Low-Rank Adaptation (LoRA) into UNet (SD1.5) and DiT (SD3) backbones. MIGLoRA is capable of preserving the base model's parameters and ensuring plug-and-play adaptability, minimizing architectural intrusion while enabling efficient fine-tuning. To support a comprehensive evaluation, we create DescripBox and DescripBox-1024, benchmarks that span diverse scenes and resolutions. The proposed method achieves state-of-the-art performance on COCO and LVIS benchmarks while maintaining parameter efficiency, demonstrating superior layout fidelity and scalability for open-world synthesis.
title Efficient Multi-Instance Generation with Janus-Pro-Dirven Prompt Parsing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.21069