Global Binary Patterns: A Novel Shape Descriptor

2013-05-23
Erdal, Sivri
Kalkan, Sinan
In this paper, we propose a novel shape descriptor, called Global Binary Patterns (GBP), based on interpreting intensity values along a direction in an image as binary numbers, converting these binary numbers to their decimal values and concatenating these decimal values as the elements of a vector that is the GBP representation of the shape. Comparing with some widelyused state-of-the-art methods in the literature, we show that GBP is very fast and its performance on several widely-used databases is comparable or better.
Citation Formats
S. Erdal and S. Kalkan, “Global Binary Patterns: A Novel Shape Descriptor,” Kyoto, Japonya, 2013, vol. 1, p. 169, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/85394.