Open AccessColor Features Based Flower Image Segmentation Using K-Means and Fuzzy C-Means
Author(s) -
Perani Rosyani,
Asep Suhendi,
Diah Hari Apriyanti,
Arya Adhyaksa Waskita
Publication year - 2021
Publication title -
building of informatics, technology and science
Language(s) - English
Resource type - Journals
eISSN - 2685-3310
pISSN - 2684-8910
DOI - 10.47065/bits.v3i3.1060
Subject(s) - artificial intelligence , segmentation , pattern recognition (psychology) , fuzzy logic , computer science , feature (linguistics) , ground truth , image segmentation , mathematics , computer vision , philosophy , linguistics
A more detail investigation of color feature for flower segmentation using K-means and fuzzy C-means was conducted in this paper. The sample images containing 1, 2, 3, 4 dianthus del- toides L flowers, obtained from ImageCLEF 2017 will be used. K-means and fuzzy C-means will use different color model components as the feature for segmenting the flower objects from their background while keeping the value of k for K-means and fuzzy C-means constant. Then the performance of the segmentation approaches will be evaluated by using the ground truth infor- mation. The evaluation parameters involved are Hausdorff distance and a number of classifier performance metrics such as accuracy, error rate, sensitivity and specivicity. It is shown that the segmentation process will greatly influenced by the use of LAB color model components