Disparity Based Image Segmentation For Occupant Classification

Abstract

Frontal "depowered" air bag systems are underdesign today to be even more effective than current air bags in saving lives, while at the same time reducing the potential of causing an air bag induced serious injury or death. Stereovision real-time occupant sensing systems (airbag suppression) have been developed at Delphi Automotive Systems 1 for use in comercial vehicle applications. One of the issues in such a system is that the irrelevant non-stationary background information within the field of view of the cameras results in less robust occupant classifications. In this paper, a disparity-based image segmentation method is provided. The input images are first segmented according to a pre-determined disparity threshold map and the real-time disparities of the occupants. Binary image processing techniques are used to reject noise introduced into the segmented images through low-resolution disparity calculations. The segmented images are then used for image feature extraction for a neural network classifier. For comparison, two neural network classifiers were created with and without the infrared image segmentation. Our experiements on segemnted images shown an increase of the classifier performance by at least 23% on a large database of IR images collected in extreme outdoor conditions.

Cost-Based Feature Transfer for Vehicle Occupant Classification

“…The first is the method of Zhang et al [46] 2 which relies on a dense Haar feature response as the feature vector and an RBF SVM for classification. Many works use Edge based features instead of Haar responses, such as [29,26,17], with variations on the classification stage. As such, the second method we compare against deploys a standard edge feature extractor (as used in [26]) followed by PCA and an RBF SVM, and we refer to this hereafter as the "Classic Approach."…”

Occupant classification for smart airbag using Bayesian filtering

“…Therefore, one focus of this work is to propose an approach that can incorporate the temporal information for classi<Ecation. Furthermore, to reduce the effect of the background scene outside the vehicle, many works adopt the sophisticated meth ods, such as adaptive threshoding [2], or stereo approaches [1], [8] to segment the region of interest (ROI) for further classi<Ecation. According to the observation, the determination of the presence of the occupant class only depends on some speci<Ec regions but not entire segmented foreground area.…”

Application of the Reeb Graph Technique to Vehicle Occupant's Head Detection in Low-resolution Range Images

“…During the last decades, many research projects have investigated solutions in the field of intelligent transportation systems (ITS) to provide cheap and reliable systems to many safety applications, including occupant classification for "smart" airbag deployment [5,7,12]. The feasi- bility of an occupant passenger classification system using low-resolution range sensor was investigated in [3].…”

“…This range sensor is advantageous since it provides directly a dense range image, independent of the illumination conditions and object textures. In [7,4], an occupant classification problem is addressed however based on low-resolution infrared images which were computed from a stereo camera system.…”