# Efficient Multi-Crop Saliency Partitioning for Automatic Image Cropping

arXiv: https://arxiv.org/abs/2506.22814 · HTML: https://mlsanity.com/paper/2506.22814/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 26 references that are in the map.
- First posted on arXiv 28 Jun 2025.

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

- [Photo Aesthetics Ranking Network with Attributes and Content Adaptation](https://mlsanity.com/paper/1606.01621/) (2016)
- [Reliable and Efficient Image Cropping: A Grid Anchor based Approach](https://mlsanity.com/paper/1904.04441/) (2019)
- [Grid Anchor based Image Cropping: A New Benchmark and An Efficient Model](https://mlsanity.com/paper/1909.08989/) (2019)
- [Learning to Compose with Professional Photographs on the Web](https://mlsanity.com/paper/1702.00503/) (2017)
- [Image Cropping with Composition and Saliency Aware Aesthetic Score Map](https://mlsanity.com/paper/1911.10492/) (2019)
- [Deep Cropping via Attention Box Prediction and Aesthetics Assessment](https://mlsanity.com/paper/1710.08014/) (2017)
- [Spatial-Semantic Collaborative Cropping for User Generated Content](https://mlsanity.com/paper/2401.08086/) (2024)
- [A2-RL: Aesthetics Aware Reinforcement Learning for Image Cropping](https://mlsanity.com/paper/1709.04595/) (2017)

## Built on by (most cited first, 1 of 1)

- [Carousel: A High-Resolution Dataset for Multi-Target Automatic Image Cropping](https://mlsanity.com/paper/2511.04680/) (2025)

## Abstract (arXiv)

Automatic image cropping aims to extract the most visually salient regions while preserving essential composition elements. Traditional saliency-aware cropping methods optimize a single bounding box, making them ineffective for applications requiring multiple disjoint crops. In this work, we extend the Fixed Aspect Ratio Cropping algorithm to efficiently extract multiple non-overlapping crops in linear time. Our approach dynamically adjusts attention thresholds and removes selected crops from consideration without recomputing the entire saliency map. We discuss qualitative results and introduce the potential for future datasets and benchmarks.