Automatic 1d histogram segmentation and application to the computation of color palettes
This article presents an implementation of the FTC (Fine-to-Coarse) algorithm for histogram segmentation, presented by Delon et al. in 2007. This algorithm uses a non-parametric a contrario approach to segment a 1D histogram into its meaningful modes. We describe also how the method may be applied t...
| Autores: | , |
|---|---|
| Tipo de documento: | artigo |
| Estado: | Versão publicada |
| Data de publicação: | 2021 |
| País: | España |
| Recursos: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositório: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/252547 |
| Acesso em linha: | http://hdl.handle.net/10261/252547 |
| Access Level: | Acceso aberto |
| Palavra-chave: | 1D histogram segmentation A contrario method Color palette |
| Resumo: | This article presents an implementation of the FTC (Fine-to-Coarse) algorithm for histogram segmentation, presented by Delon et al. in 2007. This algorithm uses a non-parametric a contrario approach to segment a 1D histogram into its meaningful modes. We describe also how the method may be applied to the hue, saturation and intensity histograms of color images in order to automatically extract their more representative colors, the so-called color palette. The algorithm for color palette extraction described in this paper is based on the one first published in 2007 by Delon et al., with an improvement that affects low-saturated colors. Several results illustrate the effectiveness of the algorithm. |
|---|