# Semantically Stable Image Composition Analysis via Saliency and Gradient Vector Flow Fusion

arXiv: https://arxiv.org/abs/2604.16500 · HTML: https://mlsanity.com/paper/2604.16500/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 18 references that are in the map.
- First posted on arXiv 14 Apr 2026.

## Builds on (most cited first, 8 of 8)

- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [DeepGaze IIE: Calibrated prediction in and out-of-domain for state-of-the-art saliency modeling](https://mlsanity.com/paper/2105.12441/) (2021)
- [Towards Artistic Image Aesthetics Assessment: a Large-scale Dataset and a New Method](https://mlsanity.com/paper/2303.15166/) (2023)
- [What is YOLOv8: An In-Depth Exploration of the Internal Features of the Next-Generation Object Detector](https://mlsanity.com/paper/2408.15857/) (2024)
- [Grid Anchor based Image Cropping: A New Benchmark and An Efficient Model](https://mlsanity.com/paper/1909.08989/) (2019)
- [Image Composition Assessment with Saliency-augmented Multi-pattern Pooling](https://mlsanity.com/paper/2104.03133/) (2021)
- [Spatial-Semantic Collaborative Cropping for User Generated Content](https://mlsanity.com/paper/2401.08086/) (2024)
- [Self-supervised Photographic Image Layout Representation Learning](https://mlsanity.com/paper/2403.03740/) (2024)

## Abstract (arXiv)

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