de
en
Schliessen
Detailsuche
Bibliotheken
Projekt
Impressum
Datenschutz
zum Inhalt
Detailsuche
Schnellsuche:
OK
Ergebnisliste
Titel
Titel
Inhalt
Inhalt
Seite
Seite
Im Dokument suchen
Heracles, Martin: Vision-based prediction of human driver behavior in urban traffic environments. 2014
Inhalt
-1.066in Contents -0.533in
1 Introduction
1.1 Objective
1.2 Motivation
1.3 Challenges
1.4 Contributions
1.5 Thesis Outline
2 State of the Art
2.1 Autonomous Vehicles
2.1.1 Highway Traffic
2.1.2 Off-road Navigation
2.1.3 Urban Environments
2.2 Driving Behavior Prediction
2.2.1 Maneuver Recognition
2.2.2 Traffic Situation Models
2.2.3 Direct Correlation Learning
2.3 Visual Scene Understanding
2.3.1 Image Segmentation
2.3.2 Scene Categorization
3 Behavior Prediction Framework
3.1 Problem Formulation
3.1.1 Input Data
3.1.2 Behavior Classes
3.1.3 Image Categories
3.2 Training Procedure
3.2.1 Ground Truth Acquisition
3.2.2 Image Classifier Training
3.2.3 Training Data Balance
3.3 Application Phase
3.3.1 Combined Prediction
3.3.2 Temporal Stabilization
3.3.3 Ambiguous Situations
4 Traffic Scene Representations
4.1 Image Filter Responses
4.1.1 Oriented Edge Filters
4.1.2 Image Grid Histograms
4.1.3 Gaussian Kernel Weighting
4.2 Object Classifier Responses
4.2.1 Object Recognition
4.2.2 Scene Decomposition
4.2.3 Feature Vector Formation
4.3 Implementation Details
4.3.1 Conditional Random Field
4.3.2 Walsh-Hadamard Features
5 Performance Analysis
5.1 Video Datasets
5.1.1 Overview
5.1.2 CamVid Dataset
5.1.3 HRI Road Traffic Dataset
5.2 Segmentation Experiments
5.2.1 Comparison to State-of-the-Art
5.2.2 Stand-alone Evaluation
5.2.3 Application to HRI Streams
5.3 Behavior Prediction Experiments
5.3.1 Behavior Class Thresholds
5.3.2 Instantaneous Prediction
5.3.3 Temporal Stabilization
5.3.4 Confidence Measures
5.3.5 Weather Recognition
6 Conclusion
References