# ShutterMuse: Capture-Time Photography Guidance with MLLMs

Source: [ShutterMuse: Capture-Time Photography Guidance with MLLMs](<https://www.sota2.com/research/paper/6a3cb624206542251d7ca624>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-28T06:00:01.087Z

Authors: Jiayu Li, Yixiao Fang, Tianyu Hu, Wei Cheng, Ping Huang, Zheheng Fan, Gang Yu, Xingjun Ma

Published: 2026-06-24T12:37:56.000Z

## Abstract

Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject-side recommendations, leaving the capture-time guidance capabilities of multimodal large language models (MLLMs) underexplored. To address this gap, we introduce CaptureGuide-Bench, a benchmark with two complementary tasks: photographer-side composition decision and refinement, and subject-side scene-conditioned pose recommendation. Our evaluation reveals limitations: general-purpose MLLMs can make composition decisions but lack precise refinement localization, while specialized aesthetic cropping models localize crops effectively but are limited to refinement; neither provides actionable pose guidance. To support model development, we further construct CaptureGuide-Dataset, comprising 130K samples with textual rationales and structured visual annotations, and develop ShutterMuse, a unified MLLM trained with supervised and reinforcement fine-tuning. Experiments on CaptureGuide-Bench show that ShutterMuse achieves the best overall photographer-side performance among evaluated baselines and competitive subject-side pose recommendation with substantially lower inference cost, demonstrating the potential of MLLMs as interactive assistants for photography during image capture.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Photographer-side guidance on CaptureGuide-Bench photographer-side guidance](<https://www.sota2.com/research/sota/photographer-side-guidance-on-captureguide-bench-photographer-side-guidance.md>): Photographer-side guidance; CaptureGuide-Bench photographer-side guidance; IoU: 74.3.

- [Human Pose Recommendation on CaptureGuide-Bench subject-side guidance](<https://www.sota2.com/research/sota/human-pose-recommendation-on-captureguide-bench-subject-side-guidance.md>): Human Pose Recommendation; CaptureGuide-Bench subject-side guidance; Plausibility: 58.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2606.25763>)

[PDF](<https://arxiv.org/pdf/2606.25763.pdf>)

[Implementation](<https://github.com/lijayuTnT/ShutterMuse>)