Monitoring time dependent image processes for detecting shifts in pixel intensities
- The problem of shift detection in image processes is addressed in this study. It is assumed that the pixel intensities follow a spatial autoregressive process, and potential shifts manifest in average intensities. The objective is to detect shifts as quickly as possible after their occurrence. To accommodate high-resolution images, a scalable technique is suggested, focusing on the surveillance of regions of interest. For shift detection, multivariate exponentially weighted moving average (EMWA) control schemes and various types of control statistics are employed. The efficiency of the proposed unique approach is demonstrated through an extensive simulation study. Additionally, recommendations for practitioners are provided regarding the selection of the chart, its setup, and calibration.
Author: | Yarema OkhrinORCiDGND, Viktoriia Petruk, Wolfgang Schmid |
---|---|
Frontdoor URL | https://opus.bibliothek.uni-augsburg.de/opus4/123510 |
ISSN: | 0943-4062OPAC |
ISSN: | 1613-9658OPAC |
Parent Title (English): | Computational Statistics |
Publisher: | Springer Science and Business Media LLC |
Type: | Article |
Language: | English |
Year of first Publication: | 2025 |
Publishing Institution: | Universität Augsburg |
Release Date: | 2025/07/16 |
DOI: | https://doi.org/10.1007/s00180-025-01645-y |
Institutes: | Wirtschaftswissenschaftliche Fakultät |
Wirtschaftswissenschaftliche Fakultät / Institut für Statistik und mathematische Wirtschaftstheorie | |
Wirtschaftswissenschaftliche Fakultät / Institut für Statistik und mathematische Wirtschaftstheorie / Lehrstuhl für Statistik | |
Dewey Decimal Classification: | 3 Sozialwissenschaften / 33 Wirtschaft / 330 Wirtschaft |
Latest Publications (not yet published in print): | Aktuelle Publikationen (noch nicht gedruckt erschienen) |
Licence (German): | ![]() |