Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/72559
Title: Application of scene labeling in auto driving
Authors: Zhan, Fangneng
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2017
Abstract: An autonomous car is a vehicle that is capable of sensing its environment and navigating without human inputs. Besides, another crucial capability of the autonomous car is about gaining a comprehensive understanding of traffic situations on a semantic level. This enables an autonomous system to distinguish from and react to objects of relevant semantic classes, including different types of vehicles, pedestrians, traffic signals as well as other types of road infrastructure. Semantic scene labeling approaches are very effective in modern camera-based perception systems. So the development and application of scene labeling may have a great influence in the field of auto driving. This thesis will give an empirical study of the latest machine learning algorithms for semantic segmentation as the basis for scene labeling. It includes a short review of the principle, structure and design of deep learning method for semantic segmentation, namely the convolutional neural networks approach. Then it will explore the application potential of deep learning method in segmentation, and evaluate the effect of different kinds of deep learning models such as different structures or parameters using a dataset. Finally, the conclusion and recommendations about the further research will be given.
URI: http://hdl.handle.net/10356/72559
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
ZhanFangneng_2017.pdf
  Restricted Access
1.79 MBAdobe PDFView/Open

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.