About session

Duplicate records are common in all digital systems that rely on collecting personal identifiers such as names, telephone numbers or scanning / distributing QR codes. Fuzzy matching algorithms for deduplicating the data are currently the norm, however they often rely on a manual, time-consuming adjudication process which is still prone to error. Matching biometric data can significantly speed up the process and bring another level of accuracy to any digital system. During this session, we will cover the technical details of biometric deduplication followed by practical information about how to set thresholds and apply different visualization techniques such as clustering.