Fingerprint Recognition via AI-Enhanced Sparse Sensor Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing touch displays with fingerprint recognition systems face challenges in capturing high-resolution fingerprint images, particularly in applications like transparent displays where sensor density is limited, making it difficult to recognize users effectively.
Innovation Solution
A fingerprint recognition method that uses a fingerprint sensor with multiple sensing units to detect and calculate geometric center points and positions of low-resolution fingerprint images, filling signals into a pixel array and using an artificial intelligence engine to generate a high-resolution candidate fingerprint image for user recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution fingerprint sensors are used, then fingerprint recognition accuracy is improved, but manufacturing cost and device complexity increase
Solution Approach 1:
The patent divides the fingerprint sensing task into two stages: first capturing low-resolution fingerprint images using sparse sensing units, then using AI algorithms to synthesize high-resolution images by filling in missing details. This segmentation allows the physical sensor to be simple while the computational processing achieves high precision.
Solution Approach 2:
The patent introduces an artificial intelligence engine as an intermediary between the low-resolution sensor data and the final high-resolution fingerprint recognition. The AI engine processes the sparse sensor inputs and generates synthetic high-resolution images, acting as a mediator that bridges the gap between limited hardware capabilities and high recognition accuracy requirements.
2Manufacturing precision
If sensor density is increased to capture high-resolution images, then image quality is improved, but transmission rate and manufacturing feasibility deteriorate
Solution Approach 1:
The patent changes the resolution parameter from the physical sensor level to the computational output level. Instead of requiring high-resolution physical sensors, the system captures low-resolution images and uses AI algorithms to generate high-resolution outputs, effectively decoupling the sensor resolution parameter from the final image quality parameter.
3Ease of manufacture
If low-density sensing units are used, then manufacturing cost and transmission rate are improved, but fingerprint recognition capability deteriorates
Solution Approach 1:
The patent replaces the mechanical solution of increasing sensor density with a computational solution using AI algorithms. Instead of adding more physical sensing units to improve recognition capability, the system uses machine learning models to synthesize high-resolution images from low-resolution inputs, substituting computational power for hardware complexity.
Data Source
AI summary
A fingerprint recognition method is provided. The method includes obtaining a plurality of fingerprint images by sensing a finger of a user, respectively calculating geometric center points corresponding to the fingerprint images, and calculating positions and offsets of the fingerprint images according to the geometric center points. The method also includes filling signals in the fingerprint images into a part of pixels in a pixel array according to the positions and the offsets of the fingerprint images, and obtaining signals of other pixels in the pixel array by inputting the signals filled in the part of pixels in the pixel array into an artificial intelligence engine. The method further includes generating a candidate fingerprint image and recognizing a user based on the candidate fingerprint image.