Projects per year
Abstract
We propose a novel optimization framework that crops a given image based on user description and aesthetics. Unlike existing image cropping methods, where one typically trains a deep network to regress to crop parameters or cropping actions, we propose to directly optimize for the cropping parameters by repurposing pre-trained networks on image captioning and aesthetic tasks, without any fine-tuning, thereby avoiding training a separate network. Specifically, we search for the best crop parameters that minimize a combined loss of the initial objectives of these networks. To make the optimization stable, we propose three strategies: (i) multi-scale bilinear sampling, (ii) annealing the scale of the crop region, therefore effectively reducing the parameter space, (iii) aggregation of multiple optimization results. Through various quantitative and qualitative evaluations, we show that our framework can produce crops that are well-aligned to intended user descriptions and aesthetically pleasing.
Original language | English |
---|---|
Article number | 108485 |
Pages (from-to) | 108485 |
Journal | Pattern Recognition |
Volume | 126 |
Early online date | 14 Feb 2022 |
DOIs | |
Publication status | Published - 7 Jun 2022 |
Keywords
- Aesthetics
- Deep network re-purposing
- Image captioning
- Image cropping
- cs.CL
- cs.CV
Fingerprint
Dive into the research topics of 'Repurposing existing deep networks for caption and aesthetic-guided image cropping'. Together they form a unique fingerprint.Projects
- 1 Finished
-
Understanding scenes and events through joint parsing, cognitive reasoning and lifelong learning (Oxford lead)
Engineering & Physical Science Research Council
1/01/16 → 28/02/22
Project: Research Councils