The widespread use of smartphones has made photography ubiquitous, yet a clear gap remains between ordinary users and professional photographers, who can identify aesthetic issues and provide actionable shooting guidance during capture. We define this capability as aesthetic guidance (AG) -- an essential but largely underexplored domain in computational aesthetics. Existing multimodal large language models (MLLMs) primarily offer overly positive feedback, failing to identify issues or provide actionable guidance. Without AG capability, they cannot effectively identify distracting regions or optimize compositional balance, thus also struggling in aesthetic cropping, which aims to refine photo composition through reframing after capture. To address this, we introduce AesGuide, the first large-scale AG dataset and benchmark with 10,748 photos annotated with aesthetic scores, analyses, and guidance. Building upon it, we propose Venus, a two-stage framework that first empowers MLLMs with AG capability through progressively complex aesthetic questions and then activates their aesthetic cropping power via CoT-based rationales. Extensive experiments show that Venus substantially improves AG capability and achieves state-of-the-art (SOTA) performance in aesthetic cropping, enabling interpretable and interactive aesthetic refinement across both stages of photo creation. Code is available at https://github.com/PKU-ICST-MIPL/Venus_CVPR2026.
Figures & tables
Figure 1: Overview of image aesthetic tasks and datasets. We follow and refine the comprehensive aesthetic task taxonomy proposed by Jin et al . [ 20 ] . A user survey of 1,069 participants shows that 91% prefer AG, a largely underexplored task with no dedicated dataset.
Figure 2: Illustration of the AGGF and examples from the proposed AesGuide dataset, showing aesthetic scores (in purple), aesthetic analyses (in orange), and aesthetic guidance (issue identification in green and shooting guidance in red) on the right side of each image.
Figure 3: Overview of the Venus framework: (1) Aesthetic guidance capability building, where AesGuide is leveraged to empower MLLMs with AG capability. (2) Aesthetic cropping power activation, which unlocks the cropping ability using CoT-based rationales.
Model
Params
Completeness
Preciseness
Relevance
Mean
Expert
Rank
Proprietary MLLMs
GPT-4o [ 16 ]
NA
0.84
1.09
1.01
0.98
1.15
8 / 7
Gemini-2.0-Pro [ 38 ]
NA
1.09
1.12
1.36
1.19
1.16
6 / 6
Qwen-VL-Max [ 2 ]
NA
0.90
1.05
0.56
0.84
0.89
11 / 11
Aesthetic MLLMs
AesExpert [ 14 ]
7B
0.33
0.56
0.51
0.47
0.56
15 / 14
Table 1: Evaluation results on the AesGuide benchmark. Mean denotes the average across three dimensions. Expert represents expert assessments, and Rank shows the order by Mean / Expert scores. Models are grouped by category: proprietary, aesthetic-oriented, and open-source general MLLMs. Best and second-best results are in bold and underlined , respectively.
Model
IoU%( ↑ )
Disp( ↓ )
R%( ↑ )
Inp.
Ina.
Specialized Aesthetic Cropping Models
ASM-Net [ 39 ]
84.86
0.0390
-
✘
✘
CACNet [ 11 ]
85.40
0.0330
-
✘
✘
HCIC [ 52 ]
85.00
0.0340
-
✘
✘
SAC-Net [ 45 ]
85.51
0.0333
-
✘
✘
UNIC [ 27 ]
84.00
0.0370
-
✘
✘
Table 2: Quantitative results on the FLMS benchmark. Inp. and Ina. denote interpretability and interactivity, respectively. Best and second-best results are in bold and underlined , respectively
Figure 4: Qualitative comparison of aesthetic cropping results among GPT-4o, AesExpert, and Venus-Q (ours) (a), along with a demonstration of Venus-Q’s interpretable and interactive aesthetic cropping capabilities (b).
Settings
Com.
Pre.
Rel.
Mean
w/o Aesthetic Analysis in Stage 1
1.09
1.17
1.41
1.22
Venus-Q (ours)
1.12
1.23
1.57
1.31
Table 3: Ablation study on the AesGuide benchmark.
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.
Aesthetic image cropping improves composition by balancing subjects, context, and image boundaries. Large vision-language models bring semantic knowledge to this task, but their scale limits lightweight photography assistance. We propose FrameLite, a compact cropping model built on Qwen3.5-0.8B. Our approach is motivated by the observation that a small model can generate better crops than its first answer suggests. We develop three-stage training to turn this potential into stronger direct predictions. Spatial and composition training establishes localization and framing preferences. Quality-Guided Online Supervision (QOS) then learns the highest-quality complete answers among the current policy's generated candidates. Greedy-Referenced Optimization (GRO) further refines the policy using positive and negative quality differences relative to its current greedy answer. Both online stages use annotation-derived feedback, while inference returns one coordinate answer without search or an external scorer. Experiments show that FrameLite outperforms the compared large-VLM croppers across public benchmarks with a substantially smaller backbone. Controlled studies confirm that selecting high-quality training answers improves the final model beyond online sampling alone. We will release our code and model weights at https://github.com/JianzeLi-114/FrameLite .
Jianze Li, Kai Liu, Ziqing Zhang +4
Shanghai Jiao Tong University · Institute of Media Technology and Experience Design, Huawei
Aesthetic image cropping aims to enhance the aesthetic quality of an image by improving its composition through spatial cropping. Previous methods often rely on saliency prediction or retrieval augmentation, ignoring the task's core requirement: a deep understanding of composition and aesthetics. Consequently, saliency-based methods struggle to make compositional trade-offs in complex scenes, while retrieval-based methods blindly refer to similar cases, lacking adaptive reasoning for unique scenes. Both approaches fail to align their automated cropping results with those of human experts. To address the above issues, we propose a novel paradigm that reformulates aesthetic cropping as a multimodal reasoning task, aiming to activate the VLM's analytical and comprehension capabilities in aesthetics. We design a Compositional Reasoning and Optimizing Preference method (CROP) that directs the VLM to think like a professional photographer. It deconstructs a complex and subjective aesthetic problem into an "analysis-proposal-decision" process, reasoning step by step through the analysis of scene elements and compositional principles. Meanwhile, our expert preference alignment module makes the model's decision consistent with human expert aesthetics. Extensive experiments across multiple datasets validate our method's superiority and component effectiveness.
Zhitong Dong, Chao Li, Jie Yu +1
Southeast University · Key Laboratory of New Generation Artificial Intelligence Technology · Alibaba Group