Fourier descriptors in cat fish identification based on teeth patches / Cecelia Assan

There are various shape identification systems available that have been used in pattern recognitions, character recognitions, and object recognitions. Shape-based identification can be divided into two categories which are boundary based such as Fourier descriptors and Chain code while region based such as various moment-based and Zernike moments. The identification of the Ariid catfish is done based on the shape of teeth patches which is located on the roof of the mouth. Segmentation plays an important role in this project. Segmentation process of the image is done based on color segmentation where it extracts the shape of the teeth patches. To enhance the extracted image of the teeth patches, morphological operation is being applied. This will assist the description of the image using Fourier descriptors. This project is focus on three genera which is Arius caelatus, Arius maculates and Osteogeneiosus militaris. Result shows that the Fourier descriptors is translation and rotation invariant.