Details

Autor(en) / Beteiligte

• Swearingen, Thomas
• Ross, Arun

Titel

Predicting Missing Demographic Information in Biometric Records Using Label Propagation Techniques

Ist Teil von

• 2016 International Conference of the Biometrics Special Interest Group (BIOSIG), 2016, p.1-5

Ort / Verlag

Gesellschaft fuer Informatik

Erscheinungsjahr

2016

Quelle

IEL

Beschreibungen/Notizen

• Biometric systems use biological attributes such as face, fingerprint, or iris to automatically recognize an individual. In many law enforcement applications, the biometric record of a person in the database is often supplemented with biographic and demographic data such as name, address, age, ethnicity, gender, etc. In such applications, some of the records may have missing or incorrect demographic data. In this work, we explore the potential of a label propagation method to impute demographic data to partially incomplete biometric records. The proposed method utilizes a graph- like structure to capture similarities between biometric records based on the available data. This structure is then used by the label propagation method to predict missing data. Experiments using the face image, name, gender and ethnicity of subjects in two datasets confirm the efficacy of the scheme in predicting missing data in biometric records.