Menerapkan K-Means Clustering untuk Segmentasi Gambar Database Berwarna

Aryo De Wibowo Muhammad Sidik (Universitas Nusa Putra)
Ilman Himawan Kusumah (Universitas Nusa Putra)
Anang Suryana (Universitas Nusa Putra)
Edwinanto (Universitas Nusa Putra)
Marina Artiyasa (Universitas Nusa Putra)
Anggy Pradiftha Junfithrana (Universitas Nusa Putra)
Yufriana Imamulhak (Universitas Nusa Putra)
Yudha Putra (Universitas Nusa Putra)

Fidelity : Jurnal Teknik Elektro
Vol 2 No 3 (2020): Edisi September 2020

Image segmentation is very important in the approach of image analysis to learn about any image. The K-means clustering technique is an algorithm widely used in image segmentation systems. This work utilizes the Lab* color space and K-means clustering to offer color-based image segmentation. This research demonstrates image segmentation of a database based on color characteristics using unsupervised K-means clustering technique implemented with MATLAB coding. The entire work is divided into two phases. Firstly, color separation augmentation from the color image database is enhanced through de-correlation stretching. Then, the six areas of the image database are clustered into three groups using the K-means clustering technique. By using the Lab* color space and K-means clustering method in the color image database, we can only show the central area of any image. We can isolate contaminated areas in medical color image databases with this approach and treat diseases quickly. We can use various approaches such as Particle swarm Optimization (PSO) for better results.