Formal evaluation strategies for feature tracking

Formal evaluation strategies for feature tracking

We present an approach for tracking space-filling features based on a two-step algorithm utilizing two graph optimization techniques. First, one-to-one assignments between successive time steps are found by a matching on a weighted, bi-partite graph. Second, events are detected by computing an indep...

Full description

Saved in:
Bibliographic Details
Published in:2016 IEEE 6th Symposium on Large Data Analysis and Visualization (LDAV) pp. 103 - 104
Main Authors: Schnorr, Andrea, Freitag, Sebastian, Helmrich, Dirk, Kuhlen, Torsten W., Hentschel, Bernd
Format: Conference Proceeding
Language:English
Published: IEEE 01-10-2016
Subjects:
Data models
Feature extraction
Image color analysis
Inspection
Manuals
Optimization
Visualization
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present an approach for tracking space-filling features based on a two-step algorithm utilizing two graph optimization techniques. First, one-to-one assignments between successive time steps are found by a matching on a weighted, bi-partite graph. Second, events are detected by computing an independent set on potential event explanations. The main objective of this work is investigating options for formal evaluation of complex feature tracking algorithms in the absence of ground truth data.
DOI:10.1109/LDAV.2016.7874339