Israeli researchers devised a new system that uses smartwatch devices which is able to verify handwritten signatures and detect forgeries.
Smartwatch System Identifies Handwritten Signatures and Forgeries
Verification technologies usually rely on dedicated digital devices, including smart pens, tablets and scanners to capture and verify signatures. A new method uses motion sensors which are available in the smartwatches. The research team notes that forgeries of handwritten signatures are made by practicing of accurately copying a person’s name by making it very difficult to distinguish between the counterfeit and the copy. The scientists developed software that use motion sensor data gathered from the movement of the wrists of the wearers. This data is used to identify the writer during the signing process. The new system uses a combination of several data sources to craft a unique signature:
Accelerometer and Gyroscope Sensors Readings.
Changes in the rotational motion and orientation.
Specific patterns of wear and use.
The researchers who have devised the new system based their hypothesis on the assumption that smartwatch owners adopt specific behavior patterns that are unique and very difficult to imitate. According to them the data can be extracted and used for such purposes.
The research study used 66 Tel Aviv University students that used a digital pen to record 15 samples of their signatures on a tablet while at the same time wearing a smartwatch. After this each student was asked to trace recordings of other subjects and imitate their signatures. The results showcased that using the collected smartwatch data their system able to verify the original with a high degree of accuracy.
The study has revealed that carrying a wrist-worn device or a fitness tracker provides much more comprehensive data than any other wearable device. This is because these types of devices carry a lot of sensors that measure the gesture of the user’s arm, hand and all fingers. All of these measures can be used to construct an offline verification method that can allow for authentication of the signature. In practice virtually any hand-worn device can be used to write and collect the handwritten signature. This includes all types of documents such as receipts, contracts and other non-digitized documents.
The research team has filed for a patent for the initial version of the system which can allow a generic smartwatch to become an accurate signature verifier. Future plans include larger-scale experimentation and the investigation of the benefits. The system can be used with digitizers to see if combining several sources can improve accuracy. The impact of the data extracted can be used in tests that can make the system useful for lie detection.
