Deteksi Level Kolesterol melalui Citra Mata Berbasis HOG dan ANN

• Main Authors: MUHAMMAD ARSYAD SIDDIK, LEDYA NOVAMIZANTI, I NYOMAN APRAZ RAMATRYANA
• Format: Article
• Language: Indonesian
• Published: Teknik Elektro Institut Teknologi Nasional Bandung 2019-05-01
• Series: Jurnal Elkomika
• Subjects: citra mata; kadar kolesterol; hog; ann.
• Online Access: https://ejurnal.itenas.ac.id/index.php/elkomika/article/view/2531
• ISSN: 2338-8323; 2459-9638

ABSTRAK Kolesterol merupakan lemak yang berada di dalam darah yang dibutuhkan untuk pembentukan hormon dan sel baru. Kadar kolesterol normal harus kurang dari 200 mg/dL, namun jika di atas 240 mg/dL akan berisiko tinggi terkena penyakit stroke dan jantung koroner. Penelitian ini menghasilkan suatu sistem yang dapat mendeteksi kadar kolesterol seseorang melalui citra mata menggunakan metode iridologi dan image processing. Citra mata diperoleh dari pasien laboratorium klinik sebanyak 120 citra mata. Proses sistem diawali dengan mengolah citra mata dengan metode cropping, resize, dan segmentasi. Metode ekstaksi ciri menggunakan Histogram of Oriented Gradients (HOG), dan klasifikasi menggunakan Artificial Neural Network (ANN). Sistem dapat mendeteksi kadar kolesterol dengan tiga level klasifikasi, yaitu normal, berisiko kolesterol tinggi, dan kolesterol tinggi dengan tingkat akurasi sebesar 93% dan waktu komputasi 0,0862 detik. Kata kunci: citra mata, kadar kolesterol, Histogram of Oriented Gradients, Artificial Neural Network   ABSTRACT Cholesterol is fat in the blood that is needed for the formation of hormones and new cells. Normal cholesterol levels should be less than 200 mg / dL, but if above 240 mg / dL will be at high risk of stroke and coronary heart disease. This study produced a system that can detect a person's cholesterol levels through eye images using iridology and image processing methods. Eye images obtained from clinical laboratory patients were 120 eye images. The system process begins with processing eye images using the method of cropping, resizing, and segmentation. Feature extraction method uses Histogram of Oriented Gradients (HOG), and classification using Artificial Neural Network (ANN). The system can detect cholesterol levels with three levels of classification, namely normal, at high risk of cholesterol, and high cholesterol with an accuracy rate of 93% and computing time of 0.0862 seconds. Keywords: eye image, cholesterol level, Histogram of Oriented Gradients, Artificial Neural Network