SA-IQA: Redefining Image Quality Assessment for Spatial Aesthetics with Multi-Dimensional Rewards
This addresses a domain-specific problem for researchers and practitioners in AI-generated content by providing a new benchmark and method for evaluating interior images, though it is incremental as it builds on existing IQA frameworks.
The paper tackles the lack of systematic evaluation for interior scenes in AI-generated image quality assessment by introducing Spatial Aesthetics, a paradigm assessing layout, harmony, lighting, and distortion, and develops SA-IQA, which significantly outperforms existing methods on the new SA-BENCH benchmark.
In recent years, Image Quality Assessment (IQA) for AI-generated images (AIGI) has advanced rapidly; however, existing methods primarily target portraits and artistic images, lacking a systematic evaluation of interior scenes. We introduce Spatial Aesthetics, a paradigm that assesses the aesthetic quality of interior images along four dimensions: layout, harmony, lighting, and distortion. We construct SA-BENCH, the first benchmark for spatial aesthetics, comprising 18,000 images and 50,000 precise annotations. Employing SA-BENCH, we systematically evaluate current IQA methodologies and develop SA-IQA, through MLLM fine-tuning and a multidimensional fusion approach, as a comprehensive reward framework for assessing spatial aesthetics. We apply SA-IQA to two downstream tasks: (1) serving as a reward signal integrated with GRPO reinforcement learning to optimize the AIGC generation pipeline, and (2) Best-of-N selection to filter high-quality images and improve generation quality. Experiments indicate that SA-IQA significantly outperforms existing methods on SA-BENCH, setting a new standard for spatial aesthetics evaluation. Code and dataset will be open-sourced to advance research and applications in this domain.