On the detection of unknown locally repeating patterns in images

Authors:
Diogo Pratas;Armando J. Pinho
Affiliations:
Signal Processing Laboratory, IEETA/DETI, University of Aveiro, Aveiro, Portugal;Signal Processing Laboratory, IEETA/DETI, University of Aveiro, Aveiro, Portugal
Venue:
ICIAR'12 Proceedings of the 9th international conference on Image Analysis and Recognition - Volume Part I
Year:
2012

Citing 8
Cited 0

Text compression

Text compression
On the Length of Programs for Computing Finite Binary Sequences

Journal of the ACM (JACM)
Content-based image retrieval with the normalized information distance

Computer Vision and Image Understanding
Complexity Based Image Artifact Detection

DCC '08 Proceedings of the Data Compression Conference
Paper: Modeling by shortest data description

Automatica (Journal of IFAC)
On the Complexity of Finite Sequences

IEEE Transactions on Information Theory
Least squares quantization in PCM

IEEE Transactions on Information Theory
The similarity metric

IEEE Transactions on Information Theory

Quantified Score

Hi-index	0.00

Visualization

Abstract

Detecting unknown repeated patterns that appear multiple times in a digital image is a great challenge. We have addressed this problem in a recent work and we have shown that, using a compression based approach, it is possible to find exact repetitions. In this work, we continue this study, introducing a procedure for detecting unknown repeated patterns that occur in a close vicinity. We use finite-context modelling to pinpoint the possible locations of the repetitions, by exploring the connection between lossless image compression and image complexity. Since repetitions are associated to low complexity regions, the repeating patterns are revealed and easily detected. The experimental results show that the proposed approach provides increased ability to eliminate false positives.