Abstract:Automatic Image Cropping is a challenging task with many practical downstream applications. The task is often divided into sub-problems - generating cropping candidates, finding the visually important regions, and determining aesthetics to select the most appealing candidate. Prior approaches model one or more of these sub-problems separately, and often combine them sequentially. We propose a novel convolutional neural network (CNN) based method to crop images directly, without explicitly modeling image aesthetics, evaluating multiple crop candidates, or detecting visually salient regions. Our model is trained on a large dataset of images cropped by experienced editors and can simultaneously predict bounding boxes for multiple fixed aspect ratios. We consider the aspect ratio of the cropped image to be a critical factor that influences aesthetics. Prior approaches for automatic image cropping, did not enforce the aspect ratio of the outputs, likely due to a lack of datasets for this task. We, therefore, benchmark our method on public datasets for two related tasks - first, aesthetic image cropping without regard to aspect ratio, and second, thumbnail generation that requires fixed aspect ratio outputs, but where aesthetics are not crucial. We show that our strategy is competitive with or performs better than existing methods in both these tasks. Furthermore, our one-stage model is easier to train and significantly faster than existing two-stage or end-to-end methods for inference. We present a qualitative evaluation study, and find that our model is able to generalize to diverse images from unseen datasets and often retains compositional properties of the original images after cropping. Our results demonstrate that explicitly modeling image aesthetics or visual attention regions is not necessarily required to build a competitive image cropping algorithm.

Multi-Scale Attention Network for Image Cropping

A Deep Network Solution for Attention and Aesthetics Aware Photo Cropping

Saliency Aware Image Cropping with Latent Region Pair.

Multi-Scale and Multi-Branch Convolutional Neural Network for Retinal Image Segmentation

An End-to-End Neural Network for Image Cropping by Learning Composition from Aesthetic Photos

Learning the Relation Between Interested Objects and Aesthetic Region for Image Cropping

Context-Aware Candidates for Image Cropping

Image Cropping with Composition and Saliency Aware Aesthetic Score Map

Robust Image Cropping by Filtering Composition Irrelevant Factors

Deep Cropping via Attention Box Prediction and Aesthetics Assessment

Multi-scale attention network for image super-resolution

Automatic image cropping with a single fully convolutional network

An Experience-based Direct Generation approach to Automatic Image Cropping

Grid Anchor Based Image Cropping: A New Benchmark and An Efficient Model

Multi-Scale Attention Network for Building Extraction from High-Resolution Remote Sensing Images

Image Inpainting Based Multi-scale Gated Convolution and Attention

Clip-Based Composition-Aware Image Cropping

Joint Probability Distribution Regression for Image Cropping

Image Cropping Assisted by Modeling Inter-Patch Relations.

Spatial-Semantic Collaborative Cropping for User Generated Content

Repurposing existing deep networks for caption and aesthetic-guided image cropping