Abstract
Object detection plays an important role in automatic driving systems. Considering the characteristics of classical and deep learning algorithms, a fusion logic is proposed to combine the advantages of these two kinds of object detectors. The relationship of detection performance among different detectors is established theoretically. According to the established theoretical relationship, the improvement of detection performance by fusion is further studied numerically. Furthermore, an optimization method is proposed to guide the design of the sub-detectors to achieve a better comprehensive performance. The effectiveness of this combined approach is validated by application to the detection of pedestrian, in which a support vector machine trained by the HOG feature of pedestrian is adopted as the classical detector and a comparatively simple transfer convolutional neural network (CNN) based on AlexNet structure acts as the deep learning detector. Several comparative tests with the classical and CNN detectors on the training dataset and other totally different dataset have been conducted to show the advantage of the combined one in ensuring detection performance with simpler network and adaptability to new application conditions.
Highlights
The detection of object is very vital for automatic driving
The rest of paper is organized as follows: the combined object detection method is presented in Section II; its performance is analyzed theoretically in Section III and the numerical optimization of this combined detector is introduced in Section IV; the effectiveness is validated by application to pedestrian detection in Section V and Section VI concludes the paper
WORK This study proposes a combined strategy for object detection to enhance the comprehensive performance by using classical and deep learning algorithms
Summary
The detection of object is very vital for automatic driving. Over the past decade, there have been significant researches on this area [1], including the detection of pedestrian [2], traffic sign [3], [4], vehicle [5], traffic light [6] and etc. (b) It is very difficult to understand the learned features They may be some public attributes of positive samples but not the object, and so when such detection network is applied to a totally new condition, its performance will be degraded greatly;. Much more computing and training resources are consumed, which is critical for on-board processors because of the limited energy and severe working environment Compared with these disadvantages, the known features with explicit physical meaning are used by the classical method to detect object. The main contributions of this paper are as follows: (1) A combined object detection method by fusing classical and deep learning methods is proposed to improve the comprehensive performance of object detector. The rest of paper is organized as follows: the combined object detection method is presented in Section II; its performance is analyzed theoretically in Section III and the numerical optimization of this combined detector is introduced in Section IV; the effectiveness is validated by application to pedestrian detection in Section V and Section VI concludes the paper
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.