Classification-driven Active Contour for Dress Segmentation

Lixuan Yang (1) , Helena Rodriguez (2) , Michel Crucianu (1) , Marin Ferecatu (1)

1 CEDRIC - VERTIGO - CEDRIC. Données complexes, apprentissage et représentations
2 Auteur indépendant

In this work we propose a a dedicated object extractor for dress segmentation in fashion images by combining local information with a prior learning. First, a person detector is applied to localize sites in the image that are likely to contain the object. Then, an intra-image two-stage learning process is developed to roughly separate foreground pixels from the background. Finally, the object is finely segmented by employing an active contour algorithm that takes into account the previous segmentation and injects specific knowledge about local curvature in the energy function. The method is validated on a database of manually segmented images. We show examples of both successful segmentation and difficult cases. We quantitatively analyze each component and compare with the well-known GrabCut foreground extraction method.

Mots clés: Dress Extraction; Clothing Extraction; Segmentation; SVM Classification; Active Contour

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]

Lixuan Yang, Helena Rodriguez, Michel Crucianu, Marin Ferecatu. Classification-driven Active Contour for Dress Segmentation. International Conference on Computer Vision Theory and Applications, Feb 2016, Rome, France. pp.22-29, ⟨10.5220/0005721000220029⟩. ⟨hal-02435274⟩