Photography Perspective Composition: Towards Aesthetic Perspective Recommendation

2025-05-27

AI Summary

Traditional photographic composition relies heavily on 2D cropping, yielding suboptimal results when subject spatial layout is poor. Professional photographers instead adjust 3D perspective to optimize 2D projection relationships—a technique known as Photographic Perspective Composition (PPC). However, PPC has long been hindered by the absence of annotated datasets and objective, quantifiable metrics for viewpoint quality assessment. To address this, we propose the first real-scene-oriented PPC paradigm: (1) we construct the first expert-annotated PPC dataset; (2) we design a Perspective Quality Assessment (PQA) network that operates without camera trajectory or textual prompts; and (3) we develop an end-to-end viewpoint recommendation framework. Experiments demonstrate that our method significantly outperforms 2D cropping baselines across multiple compositional aesthetics metrics while preserving scene geometric fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Constraint Satisfaction and Optimization: Constraint Programming
• Knowledge Representation and Reasoning: Qualitative Reasoning

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content

Abstract

Traditional photography composition approaches are dominated by 2D cropping-based methods. However, these methods fall short when scenes contain poorly arranged subjects. Professional photographers often employ perspective adjustment as a form of 3D recomposition, modifying the projected 2D relationships between subjects while maintaining their actual spatial positions to achieve better compositional balance. Inspired by this artistic practice, we propose photography perspective composition (PPC), extending beyond traditional cropping-based methods. However, implementing the PPC faces significant challenges: the scarcity of perspective transformation datasets and undefined assessment criteria for perspective quality. To address these challenges, we present three key contributions: (1) An automated framework for building PPC datasets through expert photographs. (2) A video generation approach that demonstrates the transformation process from suboptimal to optimal perspectives. (3) A perspective quality assessment (PQA) model constructed based on human performance. Our approach is concise and requires no additional prompt instructions or camera trajectories, helping and guiding ordinary users to enhance their composition skills.

Problem

• Overcoming limitations of 2D cropping in photography composition
• Addressing lack of datasets for perspective transformation
• Establishing criteria for assessing perspective quality

Innovation

• Automated framework for PPC dataset creation
• Video generation for perspective transformation visualization
• Human-based perspective quality assessment model