Vehicle Cabin Face-Image Segmentation for Seat Row Determination

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing occupant detection systems in vehicles face challenges in accurately determining the front-and-rear relationship between seated occupants from captured images, which is crucial for personalized vehicle settings like air conditioning adjustments.

Innovation Solution

An information processing apparatus that detects face images within a vehicle cabin and determines the front-and-rear relationship by comparing the areas of detected face images with a determination threshold value, dividing the image into segments corresponding to driver and passenger seats, and using correction factors for gender, age, and orientation to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If face images are detected from captured images to determine seating positions, then occupant detection capability is provided, but determination accuracy of front-and-rear relationship between occupants is insufficient

Engineering Contradiction:
Improvedetermination accuracy of front-and-rear relationshipVSAvoidreliability of seating arrangement determination
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The captured image is divided into multiple segments corresponding to different seating rows. By segmenting the image space according to the longitudinal direction of the vehicle, the system can determine which segment each detected face image belongs to, thereby establishing the front-and-rear relationship between occupants based on their respective segment positions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension for determination by comparing the areas of detected face images with a determination threshold value. This area-based dimensional information, combined with segment positioning, enables more accurate judgment of front-and-rear relationships between occupants.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If simple face detection is used, then detection speed is maintained, but accuracy in determining seating arrangement is insufficient

Engineering Contradiction:
Improveseating arrangement determination accuracyVSAvoidcomplexity of determination process
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The image is segmented into multiple regions corresponding to different seating rows. This segmentation approach maintains relatively simple processing by dividing the determination task into distinct spatial zones, while significantly improving accuracy through the additional spatial dimension.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent utilizes face image area as an additional parameter for determination. By comparing the area of detected face images with a determination threshold and combining this with segment position information, the system enhances determination accuracy without requiring complex processing algorithms.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3869394B1Information processing apparatus, program, and information processing method
Publication Date: 2024.04.17 FAURECIA CLARION ELECTRONICS CO LTD
  • EP3869394B1 patent drawingFigure 1
  • EP3869394B1 patent drawingFigure 2
  • EP3869394B1 patent drawingFigure 3

AI summary

Provided is an information processing apparatus that is to enhance determination accuracy in determining a front-and-rear relationship between positions where occupants are seated, from a captured image acquired by capturing an inside of a vehicle cabin. A server apparatus 100 includes: a detection section 131 that detects a face image of an occupant riding in a vehicle 3, from a captured image of an inside of a vehicle cabin of the vehicle 3; and a determination section 133 that, when a plurality of face images are detected, determines a front-and-rear relationship between positions where respective occupants corresponding to the face images are seated, by comparing each of areas of the detected face images with a determination threshold value, wherein the determination section 133 divides the captured image, in a direction parallel to a longitudinal direction of the vehicle 3, into a first segment 31 including a driver seat and a second segment 33 including a front passenger seat and, when the plurality of face images are detected in the first segment 31 or the second segment 33, determines the front-and-rear relationship between the positions where the respective occupants corresponding to the face images are seated, by comparing each of the areas of the face images with the determination threshold value.