Universum Learning for Semi-Supervised Signature Recognition from Spatio-Temporal Data

Abstract : We present a novel approach towards signature recognition from spatio-temporal data. The data is obtained by recording gyroscope and accelerometer measurements from an embedded pen device. The idea of Universum learning was previously presented by Vapnik and recently popularized in machine learning community. It assumes that the decision boundary of a classifier lies close to data with high uncertainty. The quality of the final classifier strongly depends on a way how to choose the Universum data and also on the representation of original data. In our paper we use a novel approach of Universum learning to classify signature data, also we present our novel idea how to sample the Universum data. At last, we also find more effective representation of the signature data itself compared to the baseline method. These three novelties allow us to outperform previously published results by 4.89% / 5.58%.
Document type :
Conference papers
Complete list of metadatas

Cited literature [18 references]  Display  Hide  Download

https://hal.univ-antilles.fr/hal-01165925
Contributor : Philippe Simon <>
Submitted on : Saturday, June 20, 2015 - 8:24:58 PM
Last modification on : Monday, June 22, 2015 - 2:02:42 PM
Long-term archiving on: Tuesday, September 15, 2015 - 8:03:15 PM

File

IGS_2015_submission_38.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01165925, version 1

Collections

Citation

Lukas Tencer, Marta Režnáková, Mohamed Cheriet. Universum Learning for Semi-Supervised Signature Recognition from Spatio-Temporal Data. 17th Biennial Conference of the International Graphonomics Society, International Graphonomics Society (IGS); Université des Antilles (UA), Jun 2015, Pointe-à-Pitre, Guadeloupe. ⟨hal-01165925⟩

Share

Metrics

Record views

146

Files downloads

226