Semantically Stable Image Composition Analysis via Saliency and Gradient Vector Flow Fusion
For researchers in computational aesthetics, this work provides a novel low-level representation for composition analysis that is robust to semantic content, though the improvement is incremental over self-supervised features.
The paper proposes VFCNet, a method for assessing photographic composition by fusing saliency and edge information into a gradient vector flow field. It achieves state-of-the-art on the PICD benchmark, with CDA-1 of 0.683 and CDA-2 of 0.629, improving by 33.1% and 36.1% over previous best.
The reliable computational assessment of photographic composition requires features that are discriminative of spatial layout yet robust to semantic content. This paper proposes a low-level representation grounded in the assumption that composition can be understood as the flow of visual attention across geometric structure. We introduce VFCNet, which fuses saliency and edge information into a gradient vector flow (GVF) field. The model computes dual-stream GVF representations, integrates them via attention, and extracts multi-scale flow features with a DINOv3 backbone. VFCNet achieves state-of-the-art performance on the PICD benchmark (CDA-1: 0.683, CDA-2: 0.629), improving by 33.1\% and 36.1\% over the previous best method. We also show that a simple classifier on self-supervised DINOv3 features substantially outperforms more sophisticated, composition-specialized models. Code is available at https://github.com/ADadras/VFCNet