Vehicle Pedestrian Detection Using Head Cues Under Occlusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer vision techniques for autonomous vehicles face inaccuracies in detecting and classifying pedestrians due to variations in height, width, pose, lighting, scale, and occlusion, leading to reduced safety and efficiency in navigation.
Innovation Solution
The use of head detection techniques, employing facial detection algorithms like neural networks, to identify pedestrians by detecting heads or facial features in image data, which improves pedestrian detection accuracy without the need for increased training data or computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional computer vision techniques are used for pedestrian detection, then the system can process general object detection, but the detection accuracy decreases due to variations in pedestrian appearance, pose, lighting, and occlusion
Solution Approach 1:
The patent segments the pedestrian detection task into two distinct stages: head detection using a specialized head detector, and full pedestrian detection using a general object detector. This segmentation allows each detector to be optimized for its specific task, with the head detector focusing on the consistent head region regardless of body variations, thereby improving overall detection accuracy and reliability under various conditions.
2Measurement precision
If more training data and computational resources are allocated to improve pedestrian detection accuracy, then detection performance may improve, but system complexity and resource requirements increase
Solution Approach 1:
The patent introduces head detection as an intermediary step between raw image input and final pedestrian detection output. The head detector serves as a mediator that provides refined positional information about the pedestrian's head, which then guides the object detector to focus on the correct region. This intermediary approach improves detection accuracy without requiring significant increases in training data or computational resources.
Data Source
AI summary
Techniques described herein relate to using head detection to improve pedestrian detection. In an example, a head can be detected in sensor data received from a sensor associated with a vehicle using a machine learned model. Based at least partly on detecting the head in the sensor data, a pedestrian can be determined to be present in an environment within which the vehicle is positioned. In an example, an indication of the pedestrian can be provided to at least one system of the vehicle, for instance, for use by the at least one system to make a determination associated with controlling the vehicle.