Efficient and Effective Detection of Repeated Pattern from Fronto-Parallel Images with Unknown Visual Contents
Signals · 24.01.2025 · 6 (1) · 4
DOI: 10.3390/signals6010004 ↗Kurz gefasst
Fachartikel in der Zeitschrift Signals, online erschienen am 24.01.2025. Verfasst von Hong Qu und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research Grants Council of the Hong Kong Special Administrative Region, China, Hong Kong Polytechnic University.
Abstract (Originalsprache)
The effective detection of repeated patterns from inputs of unknown fronto-parallel images is an important computer vision task that supports many real-world applications, such as image retrieval, synthesis, and texture analysis. A repeated pattern is defined as the smallest unit capable of tiling the entire image, representing its primary structural and visual information. In this paper, a hybrid method is proposed, overcoming the drawbacks of both traditional and existing deep learning-based approaches. The new method leverages deep features from a pre-trained Convolutional Neural Network (CNN) to estimate initial repeated pattern sizes and refines them using a dynamic autocorrelation algorithm. Comprehensive experiments are conducted on a new dataset of fronto-parallel textile images as well as another set of real-world non-textile images to demonstrate the superiority of the proposed method. The accuracy of the proposed method is 67.3%, which represents 20% higher than the baseline method, and the time cost is only 11% of the baseline. The proposed method has been applied and contributed to textile design, and it can be adapted to other applications.
Förderung
- Research Grants Council of the Hong Kong Special Administrative Region, China
- Hong Kong Polytechnic University
Zitieren
APA
Qu, H., Zhou, Y., Mok, P. Y., Flatz, G., & Li, L. (2025). Efficient and Effective Detection of Repeated Pattern from Fronto-Parallel Images with Unknown Visual Contents. Signals, 6(1), 4. https://doi.org/10.3390/signals6010004
Vancouver
Qu H, Zhou Y, Mok PY, Flatz G, Li L. Efficient and Effective Detection of Repeated Pattern from Fronto-Parallel Images with Unknown Visual Contents. Signals. 2025;6(1):4. doi:10.3390/signals6010004
BibTeX
@article{Qu2025,
author = {Qu, Hong and Zhou, Yanghong and Mok, P. Y. and Flatz, Gerhard and Li, Li},
title = {Efficient and Effective Detection of Repeated Pattern from Fronto-Parallel Images with Unknown Visual Contents},
journal = {Signals},
year = {2025},
volume = {6},
number = {1},
pages = {4},
doi = {10.3390/signals6010004}
}
EUVIMED