DRIVER MONITORING USING SPARSE REPRESENTATION WITH PART-BASED TEMPORAL FACE DESCRIPTORS

Abstract

Many driver monitoring systems (DMSs) have been proposed to reduce the risk of human-caused accidents. Traditional DMSs focus on detecting specific predefined abnormal driving behaviors, such as drowsiness or distracted driving, using generic models trained with the data collected during abnormal driving. However, it is difficult to collect sufficient representative training data to construct generic detection models, which are applicable to all drivers. Consequently, this system proposes a new personal-based hierarchical DMS (HDMS). During driving, the first layer of the proposed HDMS detects normal and abnormal driving behavior based on normal personal driving models represented by sparse representations. When abnormal driving behavior is detected, the second layer of the HDMS further determines whether the behavior is drowsy driving behavior or distracted driving behavior. The experimental results obtained for three datasets show that the proposed HDMS outperforms existing state-of-the-art DMS methods in detecting normal driving behavior, drowsy driving behavior, and distracted driving behavior.

Existing System

Multistage spatial temporal network, SVM algorithm

Proposed System

This paper proposes a novel personalized hierarchical DMS (HDMS) to efficiently and effectively detect normal and abnormal driving behaviors. The proposed HDMS contains two layers. The first layer aims to detect normal driving behavior using the part-based temporal face descriptors extracted from the driver’s testing video. If the descriptors are similar to the normal PDMs constructing using the part based temporal face descriptors previously extracted from the driver’s personal training video, the HDMS infers directly that the driver is in the normal driving state. In other words, it is unnecessary to examine the descriptors using multiple abnormal driving behavior models, and hence the computation time is significantly reduced. However, if the descriptors are dissimilar to those in the normal PDMs, the system infers that the driver is in an abnormal driving state, and hence the descriptors are further examined in the second layer of the HDMS in order to determine whether the driver exhibits drowsy driving behavior or distracted driving behavior. If the descriptors are dissimilar to both the drowsy PDMs and the distracted PDMs, the system infers the existence of some other (unspecified) form of abnormal driving behavior. 

Architecture

BLOCK DIAGRAM