A novel way of pedestrian detection using neural network with a weighted fuzzy membership function
- Authors
- Qu, L.; Lim, J.S.
- Issue Date
- Nov-2016
- Publisher
- American Scientific Publishers
- Keywords
- HOG; INRIA Dataset; NEWFM; Pedestrian detection; SVM
- Citation
- Advanced Science Letters, v.22, no.11, pp.3516 - 3519
- Journal Title
- Advanced Science Letters
- Volume
- 22
- Number
- 11
- Start Page
- 3516
- End Page
- 3519
- URI
- https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/8873
- DOI
- 10.1166/asl.2016.7866
- ISSN
- 1936-6612
- Abstract
- Pedestrian detection is a very important part of artificial intelligence and computer vision. The normal ways to accomplish pedestrian detection include HOG (histograms of oriented gradient), Haar-like, and some other descriptors with SVM (support vector machine) or AdaBoost classifiers. Because of the lack of new classifiers and progress of neural networks on classification area, neural network can be a good classifier in the field of pedestrian detection. In this paper, we study a novel classifier NEWFM (Neural Network with a Weighted Fuzzy Membership Function) by using the HOG and Haar-like descriptors. We use the INRIA data set. We use NEWFM for the learning part and detection and compare the traditional methods of pedestrian detection to evaluate performances. The result shows that the NEWFM as new classifiers have better performance than the old ones. © 2016 American Scientific Publishers. All rights reserved.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - IT융합대학 > 컴퓨터공학과 > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/8873)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.