Face Mask Detection via Facial Zone Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional image recognition systems are unreliable and inefficient in determining whether individuals are wearing face masks, particularly in controlled areas, due to the indefinite shapes of masks and limitations in artificial intelligence technology, leading to ineffective access control and increased infections.
Innovation Solution
A computer-implemented face mask detection method and system that involves obtaining model images of users wearing masks, authenticating these images, and comparing them with investigation images to determine the protection status, generating output signals for access control or notifications based on whether the user is wearing a mask.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image recognition systems are used to detect face masks, then the system complexity remains low, but the detection reliability is poor due to indefinite mask shapes and AI limitations
Solution Approach 1:
The patent segments the face detection task into distinct zones: a first zone for the mouth area and a second zone for the nose area. By dividing the detection region and processing each zone separately with zone-specific parameters, the system achieves reliable mask detection without requiring overly complex global processing, thus resolving the contradiction between reliability and complexity.
Solution Approach 2:
The patent applies local quality by using different detection parameters and processing approaches for different facial zones. The mouth zone and nose zone are detected with different criteria, allowing the system to adapt to the indefinite shapes of various mask types while maintaining detection reliability without uniform complex processing across the entire face.
2Measurement precision
If traditional access control systems are used, then the device complexity is low, but the ability to differentiate between masked and unmasked users is insufficient
Solution Approach 1:
The patent implements dynamic detection by adjusting detection parameters and processing approaches based on the specific characteristics of each facial zone and the detected features. The system dynamically adapts its measurement criteria for mouth and nose zones, achieving high precision in mask compliance detection while avoiding the need for a completely complex access control system architecture.
Solution Approach 2:
The patent performs preliminary face detection and zone identification before applying mask detection algorithms. By pre-segmenting the face into detection zones and identifying key features in advance, the system achieves accurate mask compliance measurement without requiring complex real-time processing during the actual access control decision.
3Productivity
If human security guards are used for mask control, then the technology complexity is low, but the effectiveness and efficiency are insufficient
Solution Approach 1:
The patent implements self-service detection where the system automatically detects, analyzes, and determines mask compliance without human intervention. The automated system performs face detection, zone segmentation, and compliance determination independently, achieving high monitoring efficiency while keeping the system complexity manageable through algorithmic automation rather than complex hardware or manual processes.
Solution Approach 2:
The patent replaces the mechanical system of human security guards with an automated image processing system. By substituting human visual inspection and decision-making with computer-based detection algorithms that analyze facial zones and determine mask compliance, the system achieves superior efficiency while maintaining reasonable technical complexity through software-based solutions.
Data Source
AI summary
A face mask detection system for detecting whether a user is wearing a mask at any given time using simple image recognition, configured to obtain an image of the user with the face mask, validate the image of the user wearing the face mask by an administrator (different from the user) in order to confirm validity of the image, and as a subsequent step use the validated image of the user with the face mask as a model image for benchmarking purposes. The model image is preferably a head profile of the user wearing a face mask. A face mask investigation unit compares the model image to an investigation image captured by an imaging system during an investigation process to determine whether the user in the investigation image is wearing the face mask based on whether an exact match is found between the model image and the investigation image.