Bennett, James and Finlayson, Graham D. (2026) Faster Simple Tone Curves. In: 34th Color and Imaging Conference, 2026-11-02 - 2026-11-06.
Preview |
PDF (Faster Simple Tone Curves)
- Accepted Version
Available under License Creative Commons Attribution. Download (5MB) | Preview |
Abstract
Tone curves, applied to images, are an effective way to enhance their visual appeal and to make details more conspicuous. Tone curves are widely used in image editing workflows and camera pipelines. Recently, it was proposed that tone curves should be simple. A simple tone curve has either 0 or 1 inflexion points (the curve shouldn't be wiggly). However, to find a simple approximation that is closest to a non-simple target curve, the prior-art method applies quadratic programming (QP) multiple times. And, such is the high computational cost of running QP that it will not always be practically possible to calculate simple tone curves (e.g. in a camera processing pipeline). In this paper, we show how the simple tone curve optimisation can be reformulated so it runs much more quickly. First, we equivalently enforce simplicity with respect to the derivative of tone curves. In the derivative - or density domain (the derivative of a tone curve is a histogram/density) - a simple curve has 0 or 1 extrema (a single maximum or a single minimum). Second, we show how we can optimise with respect to an extrema constraint using a method called the Pool Adjacent Violators algorithm. PAVA is, both from a computational complexity and run-time perspective, a much more efficient algorithm: we can find simple curves 300 times faster. For the MIT-Adobe FiveK dataset with 25,000 tone curves, the QP and PAVA optimisations produce visually similar results. Simple tone curves are closer to camera pipeline ready.
| Item Type: | Conference or Workshop Item (Paper) |
|---|---|
| Faculty \ School: | Faculty of Science > School of Computing Sciences |
| UEA Research Groups: | Faculty of Science > Research Groups > Colour and Imaging Lab |
| Depositing User: | LivePure Connector |
| Date Deposited: | 13 Nov 2026 10:09 |
| Last Modified: | 13 Nov 2026 10:09 |
| URI: | https://ueaeprints.uea.ac.uk/id/eprint/105102 |
| DOI: |
Downloads
Downloads per month over past year
Actions (login required)
![]() |
View Item |
Tools
Tools