M ILHAM HATTA, -
PENGARUH IMAGE ENHANCEMENT CONTRAST STRETCHING
DALAM KLASIFIKASI CT-SCAN TUMOR GINJAL MENGGUNAKAN
DEEP LEARNING.
JURNAL INOVTEK POLBENG - SERI INFORMATIKA, 9 (1).
ISSN 2527-9866
Abstract
Abstract - The kidneys are essential organs in the human body, crucial for removing metabolic waste.
Impaired kidney function can lead to an irreversible decline, known as chronic kidney disease. Another
condition affecting the kidneys is kidney tumors, which are the third most common type of tumor after
prostate and bladder tumors, with around 208,500 cases (2%) of all cancer cases globally. This research
employs the Image Enhancement Contrast Stretching technique to enhance CT scan images of kidney
tumors for deep learning classification using the EfficientNet-B0 architecture. The dataset is divided
into 80% for training, 10% for validation, and 10% for testing, resulting in 1824 training, 228 validation,
and 228 test data points per class. Hyperparameters include Adamax and RAdam optimizers with
learning rates of 0.01, 0.001, and 0.0001. The highest performance was achieved using the Image
Enhancement Contrast Stretching technique with the Adamax optimizer and a learning rate of 0.01,
yielding 99.92% accuracy, 99.85% precision, 100% recall, and a 99.92% F1-score. For the original
dataset with the same optimizer and learning rate, the best performance was 99.12% accuracy, 98.28%
precision, 100% recall, and a 99.13% F1-score. This technique has proven to enhance kidney tumor
classification models.
Keywords: Classification, Contrast Stretching, Deep Learning, EfficientNet-B0, Kidney Tumor.
Actions (login required)
 |
View Item |
