Vehicle Access Face Cropping for Multi-Person Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle access systems face challenges in accurately identifying authorized individuals due to multiple persons being within the camera's field-of-view, leading to incorrect facial recognition and potential lockout scenarios.
Innovation Solution
Implementing a cropping operation on captured images to focus on the largest face or least-obstructed shoulders within the scene, combined with facial recognition algorithms to enhance accuracy, and using a lockout counter to manage repeated recognition failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If facial recognition is performed on all detected persons in the field-of-view, then comprehensive security checking is achieved, but incorrect identification occurs when multiple people are present
Solution Approach 1:
The patent segments the image processing task by first detecting multiple persons in the field-of-view, then applying cropping operations to isolate individual persons into separate image regions. This segmentation allows the facial recognition algorithm to process each person independently, preventing incorrect identification while maintaining comprehensive security checking.
Solution Approach 2:
The patent extracts the relevant facial region from the full scene by cropping the image to contain only the detected person's face. This extraction removes distracting elements and other persons from the processing area, thereby improving identification accuracy without requiring complex multi-person analysis algorithms.
2Area of stationary object
If the camera field-of-view is widened to capture more area, then more authorized persons are detected, but recognition accuracy decreases due to multiple people in the frame
Solution Approach 1:
The patent maintains a wide camera field-of-view to capture multiple persons, then segments the captured image into individual person regions through detection and cropping. This allows the system to benefit from the wide field-of-view for comprehensive detection while achieving high precision through subsequent segmentation and focused processing of each person's facial features.
Data Source
AI summary
A computer includes a processor and a memory, the memory stores instructions executable by the processor to generate a first cropped image by applying a first cropping operation to a captured image, in which the first cropping operation is selected based on a first component to be actuated. The instructions can additionally be to apply a live person image recognition algorithm to the first cropped image to determine whether a recognized person is represented in the first cropped image and to output a signal to initiate actuation of the first component in response to the live person image recognition algorithm responsive to determining that the recognized person is in the first cropped image and is permitted to actuate the first component.