Separability filter for detecting circular shape

Detection of circular objects in a given image using Fisher criterion

The zip file contains the implementation and usage examples of algorithm from [1] and [2] for detecting circular objects in a given image.
The algorithm in [2], called separability filter, computes the Fisher criterion with a circular shape of a mask filter all over the image by sliding window. From the computation of the Fisher criterion, we obtain a separability map where the local peaks are most likely the center of the circular objects. To speed up the computation of [2], the work of [1] approximates the shape of circular shape with four combined rectangular shapes and uses Integral image in the computation. For the details please refer to [1] and [2].
Please read the included readme.txt file for the description of the files.

A demonstration video can be found at https://youtu.be/NB6AZvOYxC0

References:
[1] Y. Ohkawa, C. H. Suryanto, K. Fukui, "Fast Combined Separability Filter for Detecting Circular Objects", The twelfth IAPR conference on Machine Vision Applications (MVA) pp.99-103, 2011.

[2] K. Fukui, O. Yamaguchi, "Facial feature point extraction method based on combination of shape extraction and pattern matching", Systems and Computers in Japan 29 (6), pp.49-58, 1998.

Our homepage: http://www.cvlab.cs.tsukuba.ac.jp/