Under-screen Fingerprint Image Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fingerprint images acquired using under-screen optical sensors often include noise patterns from thin film transistors (TFTs) in the display unit, which deteriorate recognition performance due to similar frequencies, making it difficult to effectively separate the user's fingerprint from the noise patterns.
Innovation Solution
A method and apparatus that convert initial fingerprint images to reveal pattern characteristics, select a pattern removal image based on similarity with pre-stored learned patterns, and process the image to remove the TFT pattern noise, using techniques like Fourier Transform and image filtering to produce a final fingerprint image for improved recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If an under-screen optical sensor is used to acquire fingerprint images, then the fingerprint image can be captured through the display unit, but the acquired image includes noise patterns from TFTs that deteriorate recognition performance
Solution Approach 1:
The patent segments the fingerprint image processing into multiple stages: initial image acquisition, pattern detection, pattern removal, and final recognition. By dividing the processing workflow, the system can selectively remove TFT patterns while preserving fingerprint features, thus resolving the contradiction between ease of operation and recognition reliability
Solution Approach 2:
The patent extracts and removes the harmful TFT patterns from the fingerprint image while retaining the useful fingerprint information. This is achieved through pattern detection algorithms that identify TFT-related noise and selectively eliminate it, thereby improving recognition performance without compromising the acquisition process
2Loss of information
If the fingerprint image includes TFT pattern noise with similar frequency, then the pattern signal is reflected from the display unit, but the pattern exists as noise that deteriorates fingerprint recognition
Solution Approach 1:
The patent converts the harmful TFT pattern noise into a detectable feature by analyzing its characteristic frequency and spatial distribution. The system uses these pattern characteristics to identify and remove the noise, transforming what was originally harmful interference into a useful indicator for noise elimination, thus preserving fingerprint information integrity
3Reliability
If pattern removal processing is applied to the fingerprint image, then the TFT pattern noise is eliminated, but additional processing steps are required
Solution Approach 1:
The patent performs preliminary pattern detection and removal before the main fingerprint recognition process. By preprocessing the image to eliminate TFT patterns in advance, the subsequent recognition algorithms work with cleaner data, improving reliability while managing complexity through staged processing rather than attempting to handle all issues simultaneously
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method effectively removes TFT pattern noise from fingerprint images, enhancing recognition performance by isolating the user's fingerprint region and improving authentication accuracy.
Implementation Method 1
a pattern corresponding to a pattern signal reflected from a thin film transistor (TFT) included in the display unit
Implementation Method 2
acquiring a converted fingerprint image by converting the initial input fingerprint image in a pre-determined manner to reveal a characteristic of a pattern
Data Source
AI summary
A method for acquiring a fingerprint image through a fingerprint image acquisition apparatus is provided. The method includes recognizing a finger of an user through a sensor included in the apparatus to generate an initial input fingerprint image when the finger contacts with a biometric recognition area of the display unit, acquiring a converted fingerprint image by converting the initial input fingerprint image in a pre-determined manner to reveal a characteristic of a pattern, included within the initial input fingerprint image, corresponding to a pattern signal reflected from at least a part of the display unit. The method further includes selecting a pattern removal fingerprint image, if a learned pattern image is stored in a memory, from the learned pattern image based on similarity. The similarity is calculated between the converted fingerprint image and a converted reference image.