SuDaCo: Surface Covert Features Data Coding

Skimming attacks on automated teller machines (ATM) are becoming more common due to easy access of the attackers to high quality skimming devices (e. g. cameras, card-readers, keyboards, etc.). One method for obtaining a user’s personal identification number (PIN) illegally is by using a counterfeit keyboard, capable of recording PINs, which covers the original keyboard in a way that ATM users do not notice its existence. To detect a covered original keyboard, different coding and authentication methods using keyboard features are investigated.

Application constraints such as the fixed camera-to-surface distance, angular camera position, medium quality camera, light variations and reflective metallic surfaces pose challenges to be tackled. At this point it must be noted, that we are not in a controllable industrial environment. An additional challenge is posed by the comparatively noisy camera-chip of the used camera. Furthermore over the course of the project the coding of the keyboards during the production turned out to be unrealisable. On this account it is necessary to develop a method that renounces on modifications of the production process.