NADS-Net: A Nimble Architecture for Driver and Seat Belt Detection via Convolutional Neural Networks

Chun, Sehyun; Ghalehjegh, Nima Hamidi; Choi, Joseph B.; Schwarz, Chris; Gaspar, John; McGehee, Daniel V.; Baek, Stephen

doi:10.1109/iccvw.2019.00295

Cited by 25 publications

(16 citation statements)

References 36 publications

(48 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The test results showed that this proposed method increased the average correct detection and reduced the training time compared to Alexnet, VGGNet-16, and GoogLeNet. Chun et al proposed a CNN model named NADS-Net (A Nimble Architecture for Driver and Seat Belt Detection via Convolutional Neural Networks) using the feature pyramid network (FPN) backbone method and multiple detection heads [Chun et al, (2019)]. The method was applied under different demographics, appearances, and illumination.…”

Section: Non-handcrafted Featurementioning

confidence: 99%

Untitled

2023

INDJCSE

View full text Add to dashboard Cite

Traffic accident fatalities are still a severe problem in the world, and one of the contributing factors remains the failure of drivers to adhere to essential traffic regulations, exemplified by their noncompliance with seat belt usage. Previous research has developed systems for detecting drivers not wearing seat belts, consisting of three subsystems: car detection, windshield detection, and seat belt detection. Several reviews for car detection have been done, but a review focusing on seat belt detection has never been carried out. Therefore, this paper aims to review seat belt detection. The initial section outlines general steps to detect drivers not wearing a seat belt. The following section compares the methods used for seat belt detection. The research concludes by describing potential methods to improve accuracy and speed.

show abstract

Section: Non-handcrafted Featurementioning

confidence: 99%

Untitled

2023

INDJCSE

View full text Add to dashboard Cite

show abstract

“…There are large data sets like Common Objects in Context (COCO) [9], which provide a large amount of key point annotations of persons in various positions. A notable attempt to process images, which utilizes the COCO data set, was proposed in [10]. Here the authors detected the seat belt with a semantic segmentation and conducted a human posture estimation by detecting the key points of a person.…”

Section: Introductionmentioning

confidence: 99%

Automated Quantification of Occupant Posture and Shoulder Belt Fit Using Safety Specific Key Points

Hartleitner

Koppisetty

Bohman

2022

IEEE Open J. Intell. Transp. Syst.

View full text Add to dashboard Cite

Virtual evaluation of automotive safety with variation in occupant posture and shoulder belt fit is gaining importance, and there is a need of methods facilitating analysis of occupant postures in driving studies. This study is aimed to develop an AI-based computer vision method to automatically quantify occupant posture and shoulder belt position over time in a car. Traceable defined key points on the occupant were related with the shoulder belt and quantified over time in real 3D coordinates by predefined key measurements, utilising the underlying spatial information of a Intel RealSense 3D Camera. The key points are defined as traceable key points relevant to relate the occupant to the vehicle environment and to estimate shoulder belt position. Key point prediction results suggest an average deviation of around 1cm per coordinate, which enable a reliable spatial categorisation of the respective tracked occupant by analysing the key measurements. This method providing continuous information of the occupant position and belt fit will be useful to identify common occupant postures as well as more extreme postures, to be used for expanding variations in postures for vehicle safety assessments.

show abstract

“…However, pose estimation networks for IVMS have not improved much. Only a few networks (Okuno et al 2018;Yuen and Trivedi 2018;Chun et al 2019;Heo et al 2020) have attempted to assess the performance in an invehicle environment, and even those have focused solely on 2D pose estimation. (Okuno et al 2018) proposed an architecture that estimated human pose and face orientation for an autonomous driving system that consisted of only three convolutional layers and a fully connected layer; through this shallow network, it can perform real-time processing.…”

Section: Introductionmentioning

confidence: 99%

“…Clothes Jacket, long-sleeve short-sleeve winter clothes 34% 33% 33% (Yuen and Trivedi 2018) suggested predicting only the arms of the driver and passengers; this network used partial affinity fields (PAF) from (Cao et al 2017). (Chun et al 2019) has the most similar architecture to our proposed network; they performed 2D pose estimation and seat-belt segmentation and used PAF to estimate 2D keypoints, but they only estimated body keypoints without face keypoints.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Integrated In-vehicle Monitoring System Using 3D Human Pose Estimation and Seat Belt Segmentation

Kim¹,

Kim²,

Kim³

et al. 2022

Preprint

View full text Add to dashboard Cite

Recently, along with interest in autonomous vehicles, the importance of monitoring systems for both drivers and passengers inside vehicles has been increasing. This paper proposes a novel in-vehicle monitoring system the combines 3D pose estimation, seat-belt segmentation, and seat-belt status classification networks. Our system outputs various information necessary for monitoring by accurately considering the data characteristics of the in-vehicle environment. Specifically, the proposed 3D pose estimation directly estimates the absolute coordinates of keypoints for a driver and passengers, and the proposed seat-belt segmentation is implemented by applying a structure based on the feature pyramid. In addition, we propose a classification task to distinguish between normal and abnormal states of wearing a seat belt using results that combine 3D pose estimation with seat-belt segmentation. These tasks can be learned simultaneously and operate in real-time. Our method was evaluated on a private dataset we newly created and annotated. The experimental results show that our method has significantly high performance that can be applied directly to real in-vehicle monitoring systems.

show abstract

NADS-Net: A Nimble Architecture for Driver and Seat Belt Detection via Convolutional Neural Networks

Cited by 25 publications

References 36 publications

Untitled

Untitled

Automated Quantification of Occupant Posture and Shoulder Belt Fit Using Safety Specific Key Points

Integrated In-vehicle Monitoring System Using 3D Human Pose Estimation and Seat Belt Segmentation

Contact Info

Product

Resources

About