SEGMENTASI LESI KULIT MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK
Total View This Week0

Institusion

Universitas Sriwijaya

Author

DEWI CHAYANTI (STUDENT ID : 09021181722070)
M. Fachrurrozi (LECTURER ID : 0222058001)
Firdaus Firdaus (LECTURER ID : 0221017801)

Subject

T58.5 General works Management information systems Cf. HD30.213 Industrial management Cf. HF5549.5.C6+ Communication in personnel management Cf. TS158.6 Automatic data collection systems (Production control) 

Datestamp

2021-07-05 07:16:38 

Abstract :
Skin lesions are the first clinical symptoms of diseases like chickenpox, melanoma and others. With digital image processing for skin cancer detection, it is possible to make a diagnosis without any physical contact with the skin. Factors such as residue (hair and ruler markers), unclear borders, variable contrast, differences in shape and color differences in dermoscopy images of skin lesions make automatic analysis quite difficult. The presence of hair on the skin lesions can be removed effectively using segmentation. Dermoscopy image segmentation has been researched and developed in many literatures using various methods. In this study, a skin lesion segmentation system was developed using the Convolutional Neural Network (CNN) method with the U-Net architecture which produced 6 results models from parameter tuning. The best model has the highest evaluation results with Pixel Accuracy, Intersection over Union (IoU), and F1 Score of 95.89%, 90.37% and 92.54%. 

File :
RAMA_55201_09021181722070.pdf
RAMA_55201_09021181722070_0222058001_0221017801_TURNITIN.pdf
RAMA_55201_09021181722070_0222058001_0221017801_01_front_ref.pdf
RAMA_55201_09021181722070_0222058001_0221017801_02.pdf
RAMA_55201_09021181722070_0222058001_0221017801_03.pdf
RAMA_55201_09021181722070_0222058001_0221017801_04.pdf
RAMA_55201_09021181722070_0222058001_0221017801_05.pdf
RAMA_55201_09021181722070_0222058001_0221017801_06_ref.pdf
RAMA_55201_09021181722070_0222058001_0221017801_07_lamp.pdf