Free-hand Character Recognition Engine for Secure PIN Entry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Visually impaired individuals face difficulties in interacting with touch screens due to the lack of haptic feedback and visual cues, making it challenging to enter sensitive information like PIN numbers securely, especially at point-of-sale terminals, which can lead to security vulnerabilities.
Innovation Solution
A POS terminal with a character recognition engine that uses free-hand drawing and a user-specific cipher, trained through an unobtrusive recognition procedure, allowing visually impaired users to draw characters in a reserved area, providing auditory cues and enhanced security through idiosyncratic encoding, making the input more secure and recognizable only with the paired user-specific engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a standard keypad display is used on the POS terminal, then security is maintained through button selection, but visually impaired users cannot operate the interface without assistance
Solution Approach 1:
The patent replaces the mechanical button selection system with a free-hand drawing interface on a touch screen. Visually impaired users draw characters in a reserved area, and the system recognizes these drawings through a character recognition engine. This substitution maintains security through user-specific cipher encoding while enabling independent operation for visually impaired users through auditory cues and haptic feedback.
2Ease of operation
If free-hand drawing is allowed in a large space on the screen, then visually impaired users can enter PIN with little guidance, but the characters become susceptible to eavesdropping via over-the-shoulder snooping
Solution Approach 1:
The patent creates a user-specific cipher that copies and encodes the drawn characters in an idiosyncratic manner. The character recognition engine is trained to recognize only the specific user's drawing patterns, creating an encrypted representation of the PIN. This copying mechanism ensures that even if someone observes the drawing, they cannot interpret the characters without the user-specific recognition model.
Solution Approach 2:
The patent applies different quality requirements to different parts of the input system. The drawing area is made large and accessible for ease of use by visually impaired users, while the character recognition engine applies strict user-specific pattern matching to ensure security. The system treats the input collection and input verification as distinct functions with different requirements.
3Measurement precision
If conventional non-user-specific character recognition engines are used, then the system can recognize standard characters, but visually impaired users draw characters that cannot be recognized
Solution Approach 1:
The patent performs preliminary training of the character recognition engine with the specific user's drawing patterns before actual PIN entry. The system collects training data from the user's free-hand drawings and uses this to create a customized recognition model. This preliminary adaptation ensures that when the user draws characters during PIN entry, the engine can accurately recognize them despite variations in drawing style.
Solution Approach 2:
The patent makes the character recognition engine dynamic and adaptive to each user's drawing patterns. Rather than using a static recognition model, the system adjusts its recognition criteria based on the specific user's handwriting characteristics. This dynamic adaptation allows the engine to accurately recognize characters drawn by the authorized user while rejecting similar drawings by others.
Data Source
AI summary
Various methods and devices that involve character recognition on a touch display, or other input device, are disclosed. Some of these may facilitate the recognition of characters that are input by people with physical impairments such as those with limited eyesight. A disclosed free-hand character recognition device comprises: an input device having stroke detection circuitry to accept input data from a user; a character recognition engine instantiated using a processing system and a data store; and a network connection to a server. The device additionally comprises a memory storing instructions to: conduct an unobtrusive user recognition procedure to obtain a user identifier from the user; download a cipher from the server to the data store using the network connection and the user identifier; and generate a prediction vector using the character recognition engine, the input data, and the cipher.