Face Authentication Material Detection Using Patterned IR Patches
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing material detection methods from images require intensive computing resources and often lack reliability, especially in complex scenarios where objects are covered by various materials, leading to potential spoofing attacks and reduced accuracy.
Innovation Solution
A method involving pattern image manipulation to generate partial images, which are processed by a data-driven model to extract material information, utilizing less training data and computational resources, thereby enhancing accuracy and preventing spoofing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If filtering techniques are used for material detection from images, then material information can be extracted, but intensive computing resources are required and reliability is reduced
Solution Approach 1:
The patent segments the full image into multiple patches and processes each patch independently through the neural network. This segmentation approach reduces the computational burden on each processing unit while maintaining detection reliability through ensemble processing of multiple patches, directly addressing the contradiction between computing resource consumption and detection reliability
Solution Approach 2:
The patent replaces traditional filtering techniques with a data-driven neural network model that has been trained to detect material properties. This substitution of the detection mechanism achieves higher reliability through learned features while the patch-based processing manages computational resources more efficiently
2Reliability
If traditional material detection methods are used, then material information can be obtained, but the methods lack reliability in complex scenarios where objects are covered by various materials
Solution Approach 1:
By dividing the image into patches, the system can handle complex scenarios where different materials are present in different regions. Each patch can be classified independently, allowing the system to adapt to varying material compositions across the object surface, thereby improving both reliability and adaptability
Solution Approach 2:
The neural network model changes its detection parameters by being trained on diverse datasets representing various materials and scenarios. This enables the model to adapt to complex situations with multiple materials by adjusting its internal parameters based on learned patterns, improving reliability without sacrificing versatility
3Measurement precision
If full images are used for material detection, then comprehensive information is available, but computing resources and training data requirements increase
Solution Approach 1:
The patent processes images as segmented patches rather than full images, which reduces the amount of data that needs to be processed and stored during training. This segmentation maintains measurement precision by allowing focused analysis of local material properties while significantly reducing the quantity of training data required
Solution Approach 2:
The patent extracts only the necessary patches from full images for training and inference, removing unnecessary portions of the image data. This extraction approach maintains material identification accuracy by focusing on relevant regions while reducing the overall data volume that needs to be processed
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 achieves accurate material identification and authentication with reduced computational expense, suitable for devices like smartphones, by using partial images and a data-driven model trained on less data, thus improving security and reliability.
Implementation Method 1
material information is derived from reflections of a pattern illumination
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention refers to a method for authorizing an object of a device to perform at least one operation on the device that requires authentication, comprising the steps: - receiving, from a user interface associated with a device, an unlock request for the device from a user; - in response to receiving the unlock request, triggering illumination of the object with patterned infrared illumination; - triggering to capture at least one image of the object using a camera located on the de- vice while the object is being illuminated with the patterned infrared illumination; - extracting material data from at least one image; and - allowing or declining the object to perform at least one operation on the device that requires authentication based on the material data.