Fingerprint Ridge Frequency Extraction via Convolutional Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Latent fingerprints of poor quality pose challenges for automatic matching systems due to low image quality and masked features, making it difficult to detect minutiae and ridges effectively.
Innovation Solution
A method involving transforming the source image into the frequency domain using a Fourier transform and applying a convolutional neural network to determine ridge frequencies, normalize the image, and extract signatures, which enhances the detection of ridge frequencies and minutiae characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If chemical and physical development techniques are used to improve latent fingerprint visibility, then the visibility of fingerprints is improved, but the image quality remains poor and backgrounds may mask the latent
Solution Approach 1:
The patent replaces traditional chemical and physical development techniques with a digital image processing approach using convolutional neural networks. The system transforms the latent fingerprint image into the frequency domain, applies deep learning algorithms to enhance ridge structures, and extracts minutiae features computationally, thereby eliminating the need for chemical developers while improving both visibility and image quality reliability
Solution Approach 2:
The patent transforms the fingerprint image from the spatial domain to the frequency domain, changing the representation parameters of the image. By operating in the frequency domain and using learned filters to enhance specific frequency components corresponding to ridge structures, the system improves visibility while maintaining image quality, overcoming the limitation of traditional development techniques
2Ease of manufacture
If traditional algorithms are used for detecting minutiae and ridges in latent fingerprints, then the detection process is simple, but the detection accuracy is low due to poor image quality
Solution Approach 1:
The patent replaces traditional algorithmic approaches with a convolutional neural network-based system. The deep learning model automatically learns optimal feature detection patterns from training data, enabling accurate detection of minutiae and ridges in low-quality latent fingerprints without requiring complex manual algorithm design, thus improving detection accuracy while maintaining ease of implementation
Solution Approach 2:
The patent applies a convolutional neural network that has been pre-trained on large datasets of fingerprint images. The network parameters are learned in advance from training data, allowing the system to perform accurate detection without requiring complex real-time processing. This preliminary training action enables the system to handle poor quality images effectively
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
This method effectively extracts signatures from low-quality fingerprint images, improving the probability of determining correct ridge frequencies and identifying individuals by comparing extracted signatures with reference fingerprints.
Implementation Method 1
transforming the source image into the frequency domain using a Fourier transform
Data Source
AI summary
A method for extracting a signature of a fingerprint shown on a source image is described. For this purpose, the source image is transformed into the frequency domain. One or more ridge frequencies are next determined by means of a convolutional neural network applied to said transformed image, n being an integer greater than or equal to 1. The source image is normalized in response to said n ridge frequencies determined. One or more signatures of said fingerprint are finally extracted from said n normalized images.