Vehicle Camera Optical Search Window Tracking via Seat Acceleration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing driver monitoring and face recognition systems in vehicles face challenges with vertical movement of cushioned vehicle seats, leading to rapid displacement of the driver's position relative to the image sensor, resulting in blurred images and increased computational effort for object tracking.

Innovation Solution

A method that measures vehicle seat acceleration and changes in vehicle acceleration in all directions, using this information to predict the driver's position and adjust the camera's optical search window, implementing optical and/or software-based image stabilization to maintain sharp images and reduce computational load.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a computationally intensive face recognition algorithm searches a complete image for a human face, then face detection accuracy is improved, but computational effort and processing time increase significantly

Engineering Contradiction:
Improveface detection accuracyVSAvoidcomputational effort
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the complete image into multiple regions of interest (ROIs) based on acceleration data prediction. Instead of searching the entire image, the algorithm focuses computational resources on predicted driver positions, segmenting the search space into relevant and irrelevant areas. This reduces the computational burden while maintaining detection accuracy within the segmented regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary action by using acceleration sensors to predict the driver's position before the actual image analysis. The motion prediction based on acceleration data pre-positions the search window, so that when face recognition is performed, the algorithm already knows where to look, eliminating the need for exhaustive full-image scanning.

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If the camera uses a fixed optical search window to track the driver, then device complexity is reduced, but tracking reliability deteriorates due to vertical seat movement

Engineering Contradiction:
Improvecamera system simplicityVSAvoidtracking reliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent applies dynamics by making the optical search window adaptive rather than fixed. The search window dynamically adjusts its position and size based on real-time acceleration data that predicts driver movement. This dynamic adjustment maintains tracking reliability during vertical seat movement while keeping the overall system relatively simple by only modifying the search window parameters.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements feedback by continuously monitoring acceleration signals and using this information to adjust the search window position. The acceleration data provides feedback about driver movement trends, which is fed back into the tracking algorithm to continuously update the predicted driver position, creating a closed-loop system that maintains reliability.

Inventive Principle:
Principle #23Feedback

3Device complexity

If image stabilization is not applied, then device complexity remains low, but image quality deteriorates due to rapid driver displacement from vertical seat movement

Engineering Contradiction:
Improvesystem complexityVSAvoidimage sharpness
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent replaces mechanical image stabilization systems with a computational approach. Instead of using moving lenses or sensors to physically stabilize the image, the system uses software-based image stabilization that processes the image data computationally. This substitution reduces mechanical complexity while achieving the same image sharpness goal through digital image processing techniques.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Reliability

If the search window size is increased to cover potential driver positions, then tracking robustness is improved, but the area to be processed increases leading to higher computational effort

Engineering Contradiction:
Improvetracking robustnessVSAvoidcomputational effort
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies local quality by concentrating computational resources on the local region where the driver is predicted to be, rather than uniformly processing a large entire search area. The acceleration-based prediction identifies a specific local region of interest, and the face recognition algorithm focuses its computational effort on this localized area with higher processing quality, while reducing or eliminating processing in other areas.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3753242B1Method for tracking an object in an optical search window of a camera for a vehicle
Publication Date: 2022.08.17 ROBERT BOSCH GMBH
  • EP3753242B1 patent drawingFigure 1~2
  • EP3753242B1 patent drawingFigure 3

AI summary

The invention relates to a method for tracking an object in an optical search window of a camera (115) for a vehicle (100), wherein the method comprises a step of inputting, in which at least one acceleration signal (130) representing an acceleration of the vehicle (100) and/or of a vehicle seat (20) is input. The method also comprises a step of determining, in which a movement vector (40) of at least one object in the vehicle interior and/or in a vehicle environment is determined using the at least one input acceleration signal (130). Finally, the method comprises a step of providing, in which a displacement vector (145) generated using the determined movement vector (140) is provided to the camera (115) for the displacement of the optical search window in an image determined by the camera (115) for detecting the object, wherein the optical search window represents a sub-region of the image input by the camera (11).