Human detection using different deep neural networks

Groefsema, Steff (2019) Human detection using different deep neural networks. Bachelor's Thesis, Artificial Intelligence.

Preview	Text AI_BA_2019_SteffGroefsema.pdf Download (3MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (119kB)

Abstract

In this thesis we compare the performance of two different deep neural networks on the task of human detection. The first network is Faster regional based convolutional neural network (Faster R-CNN), the second Single Shot Multibox Detector. The research question is: Is Faster R-CNN able to achieve a higher mean average precision in comparison to a Single Shot Multibox Detector? The hypothesis is that Faster R-CNN is able to achieve a higher mean average precision in comparison to a Single Shot Multibox Detector. We compare both networks in pairs using ResNet-50 and Inception-V2 as a backbone. Both networks are trained on a subset of the Pascal VOC 2007+2012 dataset and the performance scores are measured using the COCO metrics. Each network has been trained 10 times and the averages of each network are compared with a t-test. After conducting the experiment, the results show that Faster R-CNN was able to achieve a significantly higher mean average precision in comparison with the Single Shot Multibox Detector with ResNet-50 and Inception-V2 as a backbone.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Wiering, M.A.
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	21 Mar 2019
Last Modified:	22 Mar 2019 13:53
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/19289

Actions (login required)

View Item