Fingerprint Sensor Region Segmentation Memory Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fingerprint recognition performance in mobile devices is often inadequate due to limited sensing area and memory capacity, leading to false recognition and limited authentication of fingerprints.
Innovation Solution
A method for processing fingerprint information that involves obtaining input images from a user fingerprint, comparing them with registered images stored in memory, and classifying them based on unique and overlapping regions, with a scoring system to determine storage and grouping, while efficiently managing memory by deleting less similar images when space is low.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If more registered images are stored in memory to improve fingerprint recognition coverage, then fingerprint recognition performance is improved, but memory capacity is exceeded
Solution Approach 1:
Multiple fingerprint images that share common regional characteristics are merged into a single composite registered image. The merging process integrates overlapping regions from multiple input images to create a comprehensive reference image that covers a wider fingerprint area, thereby improving recognition performance without proportionally increasing memory storage requirements.
Solution Approach 2:
The fingerprint image is divided into multiple regions, and only unique non-overlapping regions are extracted and stored. By segmenting the fingerprint space and storing only distinctive regional features rather than complete redundant images, the system achieves better recognition coverage within limited memory capacity.
2Adaptability or versatility
If a larger sensing area is used to capture more fingerprint regions, then authentication coverage is improved, but device size increases
Solution Approach 1:
The system transitions from spatial expansion (larger physical sensing area) to dimensional expansion in data space (combining multiple partial images). By stitching together multiple smaller sensing area images into a comprehensive fingerprint map, the system achieves wide authentication coverage without increasing the physical footprint of the sensor.
3Measurement precision
If more fingerprint images are stored to handle diverse fingerprint patterns, then recognition accuracy is improved, but false recognition increases due to redundant storage
Solution Approach 1:
The system extracts and stores only the unique non-overlapping regions from fingerprint images, removing redundant overlapping parts. By identifying and retaining only the distinctive features in each image while discarding duplicate information, the system maintains high authentication accuracy without the pitfalls of redundant storage.
Data Source
AI summary
A method for processing fingerprint information may include obtaining an input image corresponding to at least a region of a user fingerprint, searching a memory storing at least one registered image for a reference registered image that matches the input image at least partially, determining a unique region of the input image in which the input image does not overlap with the reference registered image when the input image is aligned with the reference registered image, and, in response to a score of the unique region of the input image being greater than or equal to a predetermined reference score, classifying the input image as belonging to a same group as the reference registered image and storing the input image in the memory.