Accurate Binary Image Selection from Inaccurate User Input - Archive ouverte HAL Access content directly
Journal Articles Computer Graphics Forum Year : 2013

Accurate Binary Image Selection from Inaccurate User Input

(1) , (2) , (3) , (4)
1
2
3
4

Abstract

Selections are central to image editing, e.g., they are the starting point of common operations such as copy-pasting and local edits. Creating them by hand is particularly tedious and scribble-based techniques have been introduced to assist the process. By interpolating a few strokes specified by users, these methods generate precise selections. However, most of the algorithms assume a 100% accurate input, and even small inaccuracies in the scribbles often degrade the selection quality, which imposes an additional burden on users. In this paper, we propose a selection technique tolerant to input inaccuracies. We use a dense conditional random field (CRF) to robustly infer a selection from possibly inaccurate input. Further, we show that patch-based pixel similarity functions yield more precise selection than simple point-wise metrics. However, efficiently solving a dense CRF is only possible in low-dimensional Euclidean spaces, and the metrics that we use are high-dimensional and often non-Euclidean. We address this challenge by embedding pixels in a low-dimensional Euclidean space with a metric that approximates the desired similarity function. The results show that our approach performs better than previous techniques and that two options are sufficient to cover a variety of images depending on whether the objects are textured.
L'opération de sélection est essentielle en traitement d'images, bien que fastidieuse à effectuer a la main. Des techniques à base de marques ont été développées. Cependant celles ci supposent que l'utilisateur fournit des marques exactes a 100%, ce qui impose à l'utilisateur des contraintes supplémentaires. Dans ce papier, nous proposons une technique de sélection qui est robuste par rapport à la qualite des marques fournies par l'utilisateur. Nous utilisons un "conditional random field" dense pour inferrer une selection binaire, de manière robuste à partir de l'entrée utilisateur. Comme ce choix impose une dimensionalité faible des espaces de travail, ainsi qu'une metrique euclidienne, nous utilisons une réduction de dimension empirique basée sur la fonction de similarité entre pixels. Nos résultats montrent qu'a entrée égale, notre méthode surpasse les travaux existants.
Vignette du fichier
Screenshot_from_2013-01-29_09_59_46.png (225.42 Ko) Télécharger le fichier Fichier principal
Vignette du fichier
sel.pdf (2.22 Mo) Télécharger le fichier
Format : Figure, Image
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-00782232 , version 1 (29-01-2013)

Identifiers

Cite

Kartic Subr, Sylvain Paris, Cyril Soler, Jan Kautz. Accurate Binary Image Selection from Inaccurate User Input. Computer Graphics Forum, 2013, 32 (2pt1), pp.41-50. ⟨10.1111/cgf.12024⟩. ⟨hal-00782232⟩
698 View
943 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More