Estimation of Feret’s diameter from pixel coverage representation of a shape

• We propose a method for Feret’s diameter estimation from digital images.
• We derive a correction term for improving accuracy of the coverage based estimator.
• We give the upper bound of the absolute error for the proposed method.
• We evaluate the proposed method on synthetic and real images.
• The proposed method provides significant improvement over existing methods.

Feret’s diameter of a shape is a commonly used measure in shape analysis. Traditional methods for estimation of Feret’s diameter are performed on binary images and are of poor precision and accuracy. We analyze and further develop a method for estimation of Feret’s diameter that utilizes pixel coverage. We improve the accuracy of the method by proposing a correction term. We provide an expression for the upper bound of the absolute error of the estimation. We evaluate the improved method and compare with existing methods for Feret’s diameter estimation, based on both binary and coverage representations of image objects. Tests confirm increased precision and accuracy of the new method, on synthetic as well as on real images.