Paper Detail
Jiarui Guo, Haojia Wei, Yiming Zhang, Yifei Liu, Yuning Gong, Hongjie Zhang, Xue Yang, Zhihang Zhong
Virtual photography asks an agent to enter a prepared 3D scene with no preselected camera pose or reference image, infer a suitable shot from scene information and a language intent, choose executable camera parameters, and render the final photograph. Recent progress in vision-language models makes this kind of spatial agent increasingly plausible, but the task stresses two capabilities that remain hard to evaluate together: complex 3D spatial understanding and abstract aesthetic judgment. We introduce PhotoFlow, a Director-Reviewer-Reflector agent for closed-loop camera search. The Director builds a soft photographic blueprint and proposes diverse candidate cameras; the Reviewer combines rule checks, visual critique, and pairwise incumbent selection; and the Reflector converts failures into region memory, dead-zone suppression, and high-explore relocation. We also introduce VPhotoBench, a benchmark of 47 open-license Blender scenes and 141 language-conditioned photography missions spanning subject placement, relational composition, and atmosphere/style. On held-out experiments, PhotoFlow achieves the strongest external quality-alignment composite and success rate among one-shot prediction, single-chain reflection, anchor-bank selection, and random search under a six-round rendering budget. To our knowledge, this is the first work to make language-conditioned virtual photography in arbitrary Blender scenes an executable agent task, and our results show that an LLM-centered spatial agent can already produce strong photographs in a setting designed to challenge both 3D reasoning and aesthetic choice.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@article{guo2026photoflow,
title = {PhotoFlow: Agentic 3D Virtual Photography Missions},
author = {Jiarui Guo and Haojia Wei and Yiming Zhang and Yifei Liu and Yuning Gong and Hongjie Zhang and Xue Yang and Zhihang Zhong},
year = {2026},
abstract = {Virtual photography asks an agent to enter a prepared 3D scene with no preselected camera pose or reference image, infer a suitable shot from scene information and a language intent, choose executable camera parameters, and render the final photograph. Recent progress in vision-language models makes this kind of spatial agent increasingly plausible, but the task stresses two capabilities that remain hard to evaluate together: complex 3D spatial understanding and abstract aesthetic judgment. We i},
url = {https://arxiv.org/abs/2605.23771},
keywords = {cs.CV, cs.AI, cs.MA, vision-language models, spatial agent, 3D spatial understanding, aesthetic judgment, PhotoFlow, Director-Reviewer-Reflector agent, photographic blueprint, camera parameters, visual critique, region memory, dead-zone suppression, high-explore relocation, VPhotoBench, Blender scenes, language-conditioned photography, LLM-centered spatial agent, code available, huggingface daily},
eprint = {2605.23771},
archiveprefix = {arXiv},
}
{}