Koi fish classification based on HSV color space

Digital image processing is still a great demand for research. Research related to digital image processing can be components of color, texture and pattern. This study focuses on the segmentation process of the body pattern of koi. Koi fish is a fish species originating from the country of Japan are...

Full description

Saved in:
Bibliographic Details
Main Authors: Kartika, Dhian Satria Yudha, Herumurti, Darlis
Format: Conference Proceeding
Language:English
Subjects:
assessment
classification
color extraction
Digital images
Feature extraction
Fish
HSV
Image color analysis
Image segmentation
K-Means
segmentation
Support vector machines
Testing
Citations: Items that cite this one
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Digital image processing is still a great demand for research. Research related to digital image processing can be components of color, texture and pattern. This study focuses on the segmentation process of the body pattern of koi. Koi fish is a fish species originating from the country of Japan are much in demand by the people of Indonesia as diverse shades of color and a unique pattern. This study focuses on 9 koi fish that will be grouped into classes. From 9 of koi fish are 281 datasets were later processed into training data and data testing. The segmentation process becomes important to obtain high accuracy before the classification process. The proposed segmentation method using the K-Means as pre-processing. K-Means method used for the separation of the object and the background with two color features are worth 0 and 1. Results of pre-processing will be displayed on color feature is worth 1; object fish that has a value of Red, Green, Blue (RGB). The value in the subsequent feature extraction RGB colors into Hue Saturation Value (HSV). The process of using the HSV color feature extraction is proposed to obtain classification results with high accuracy values. The testing process using tools Weka 3.8.0 Classification with Naive Bayes method compared with Support Vector Machine (SVM) which both use the K-Fold Cross Validation. The test results showed the Naive Bayes without K-Fold Cross Validation and SVM using K-Fold Cross Validation together have a value of high accuracy of 97%. It can be concluded that the segmentation method using the K-Means and HSV capable of providing high accuracy impact on the testing process by 97%.
ISSN:2996-1378
DOI:10.1109/ICTS.2016.7910280