Writer Identification Using GMM Supervectors and Exemplar-SVMs. (March 2017)

• Record Type: Journal Article
• Authors: Christlein, Vincent; Bernecker, David; Hönig, Florian; Maier, Andreas; Angelopoulou, Elli
• Is Part Of: Pattern recognition. Volume 63(2017:Mar.)
• Page Start: 258
• Page End: 267
• Publication Date: 2017-03
• Subjects: Writer identification -- GMM supervectors -- Exemplar-SVM; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2016.10.005
• Languages: English
• ISSNs: 0031-3203

Abstract

Abstract: This paper describes a method for robust offline writer identification. We propose to use RootSIFT descriptors computed densely at the script contours. GMM supervectors are used as encoding method to describe the characteristic handwriting of an individual scribe. GMM supervectors are created by adapting a background model to the distribution of local feature descriptors. Finally, we propose to use Exemplar-SVMs to train a document-specific similarity measure. We evaluate the method on three publicly available datasets (ICDAR / CVL / KHATT) and show that our method sets new performance standards on all three datasets. Additionally, we compare different feature sampling strategies as well as other encoding methods. Abstract : Highlights: We provide a robust method for offline writer identification. We integrate Exemplar-SVMs, which greatly improve the recognition rate. We show that RootSIFT descriptors sampled at the contour are strong features. We investigate different encoding strategies next to GMM supervectors. We evaluate our method on three publicly available datasets (ICDAR13, CVL, KHATT).