• Aucun résultat trouvé

Robustness Assessment of Texture Features for the Segmentation of Ancient Documents

N/A
N/A
Protected

Academic year: 2021

Partager "Robustness Assessment of Texture Features for the Segmentation of Ancient Documents"

Copied!
6
0
0

Texte intégral

Loading

Figure

Fig. 1. Examples of three types of noise in semi-synthetic documents: (a) a non-noisy image, (b) two independent dark/bright spots, (c) two overlapping dark/bright spots, and (d) disconnection bright spot.
Fig. 3. Evaluation of the extracted textural features using different accuracy metrics on enhanced images (non-local means filtering and superpixel  tech-niques) and degraded images: silhouette width (SW), purity per block (PPB), precision (P), recall (R),
Fig. 5. Texture feature evaluation results on simplified ancient document images (cf. Figures 5(a), 5(f), and 5(k)), enhanced images by non-local means (cf

Références

Documents relatifs

Recent study [30–35] addressing texture-based image segmentation and classification using very high resolution (VHR) remote sensing images has proved the capacity of the local

Palm and Lehmann (Palm & Lehmann, 2002) show that the Gabor transform is competitive at describing image textures and these texture feature extraction from the color

indeed, given the reference BioCode of a genuine user, the intruder will try to generate an eligible BioCode, using different available data (token, FingerCode.. Recall that

Dimensionality and performance evaluation of each texture-based feature set for documents of the training and testing datasets in the following three cases: with full texture

The goal of this paper is to deal with these two issues, and to show that Fast2Sum and 2Sum (Algorithms 1 and 2) are much more robust than it is in general believed: for any

This article presents a framework for the texture-based seg- mentation of historical digitized book content with little a priori knowledge in order to evaluate three approaches based

La MPR a développé une approche globale du patient, en s’appuyant notamment sur la Classification internationale du fonctionnement, du handicap et de la santé (CIF), qui

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des