Betta Fish Image Classification Using Artificial Neural Networks with Gabor Extraction Features

Abstract

Betta fish also known as battling fish, is a type of freshwater fish that is well known among ornamental fish lovers. For this reason, the analyst proposes Betta Fish Picture Grouping Utilizing Artificial Neural Networks with the Color Gabor feature. The test results have 3 parameters, namely precision, recall, and accuracy. The level in comparison using a comparator between 50:50. The results obtained starting from the Gabor feature with CMYK precision color have test results reaching 37.94%. Then the recall has a value of 30.40% and accuracy in the existing accuracy reaches 56.71%. From the results of testing the Gabor feature with HSV precision color, reached 38.69%. Then the recall has value of 34.92% and accuracy in the existing accuracy reaches 54.69%. The Gabor feature with RGB precision reaching 39.40% at a 50:50. Then the recall has a value of 32.28% at a 50:50. The level of accuracy in the existing accuracy reaches 58.85% with a ratio of 50:50. From this it can be concluded that the Gabor feature with GRB color has the best accuracy value at a ratio of 50:50. The Gabor feature with RGB color is the best result in betta fish classification using Artificial Neural Networks.