Measurement of asymmetry of stripe patterns in animals

The deviation from perfect symmetry of the left and right sides of a bilateral animal is known to indicate the developmental stress of the animal. It has also been used to study the general health of a population. For animals like tigers and zebras which have natural stripes on their bodies, the asymmetry between the two sides can be measured by analysing the stripe patterns taken from images of these animals. In this paper we propose an algorithm for measuring this asymmetry. We use a Short-Time Fourier Transform based approach to measure the local frequencies at different locations and compare the resulting frequency maps. The mean absolute error between the left and right frequency maps yields a measure of the asymmetry. Our results show a surprising level of asymmetry in tiger stripes compared to zebras.