Facial Analytics System for High-Throughput Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional facial recognition systems are inefficient in high-throughput settings like airport customs and immigration, requiring subjects to pose and resulting in latency and low throughput, with no automated methods to identify individuals on terror or no-fly lists.
Innovation Solution
A method and system for facial analytics that captures and analyzes images in real-time, selecting high-quality images for facial recognition without manual interaction, using a local network for image processing and database updates, and sending only cropped facial images to a central server for recognition, while discarding non-relevant images and signaling individuals to adjust their position for optimal detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional facial recognition systems are used in high-throughput settings, then identification accuracy is maintained, but processing speed and throughput are reduced due to latency and posing requirements
Solution Approach 1:
The system performs preliminary actions by continuously capturing and buffering images before the actual recognition moment. Multiple images are pre-captured and stored in a buffer, allowing the system to quickly select and process the best quality image when recognition is needed, eliminating the need for real-time posing and reducing latency.
Solution Approach 2:
The system dynamically adjusts the image selection process by evaluating multiple captured images based on quality metrics. Instead of using a single static image, the system dynamically selects the optimal image from the buffer that best meets recognition requirements, adapting to varying capture conditions in real-time.
2Measurement precision
If multiple images are captured and analyzed for each individual, then image quality and recognition accuracy are improved, but data transmission bandwidth and storage requirements increase
Solution Approach 1:
The system extracts only the essential facial recognition data from captured images. After analyzing multiple images to determine the best quality image, the system extracts and transmits only the cropped facial image and recognition results, leaving the full-resolution source images at the local site. This significantly reduces the volume of data that needs to be transmitted while preserving recognition accuracy.
3Productivity
If automated facial recognition is implemented without posing requirements, then throughput and speed are improved, but system complexity increases due to multiple image capture and selection processes
Solution Approach 1:
The system segments the facial recognition process into distinct functional modules: image capture, buffer management, image quality analysis, selection logic, and recognition processing. Each module handles a specific aspect of the workflow, making the overall complex system manageable and maintainable. The segmentation allows parallel processing and independent optimization of each component.
4Loss of energy
If a local network is used for image processing, then bandwidth requirements are reduced, but processing capability and database access are limited compared to centralized systems
Solution Approach 1:
The system implements local quality by enabling autonomous processing capabilities at the local site. The local network can independently perform image capture, analysis, selection, and recognition without requiring constant connection to centralized servers. This local autonomy reduces bandwidth consumption while maintaining full processing capability for routine operations.
Data Source
AI summary
A method for facial analytics includes capturing a series of images of individuals from a camera into a circular buffer and selecting a plurality of images from the buffer for analysis in response to a trigger event, wherein the plurality of images are chronologically proximate before and/or after the trigger event in time. The method includes analyzing the plurality of images to determine image quality and selecting one of the plurality of images based on image quality to form a cropped facial image most likely to result in positive facial recognition matching. Methods of signaling to control the pedestrian traffic flow can maximize the individuals' facial alignment to the capturing camera's field of view. Non-relevant facial images associated with individuals outside a given region of interest can be discarded. Facial recognition is run on the resultant cropped facial image. Output can be displayed with information from the facial recognition.