Comparison of support vector machines and deep learning for vehicle detection
Küçük Resim Yok
Tarih
2018
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
CEUR-WS
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
The main goal of this paper is to compare different Vehicle Detection algorithms and to provide an effective comparison technique for developers and researchers. During this study, fine tunings are suggested to improve the implementations of these algorithms. Our focus on Support Vector Machines (SVM) and then Deep Learning based approaches. The SVM based vehicle detection implementation utilizes Histogram Oriented Gradients (HOG). The deep learning approach we consider is the YOLO implementation. Our evaluation employs 400 random frames extracted from a real world driving video. As stated by the experimental results, YOLO is more accurate with %81.9 success than SVM which only scored %57.8. © 2018 CEUR-WS.
Açıklama
3rd International Conference on Recent Trends and Applications in Computer Science and Information Technology, RTA-CSIT 2018 -- 23 November 2018 through 24 November 2018 -- -- 143396
Anahtar Kelimeler
Kaynak
CEUR Workshop Proceedings
WoS Q Değeri
Scopus Q Değeri
N/A
Cilt
2280