Skripsi
PERBAIKAN KUALITAS CITRA ENDOSKOPI MENGGUNAKAN DISCRETE WAVELET TRANSFORM DAN UNSHARP MASKING UNTUK SEGMENTASI INSTAN DENGAN YOLOV8-SEG
Colorectal cancer is a colon cancer characterized by the presence of colorectal polyps. The detection of colorectal polyps can be performed automatically using deep learning methods based on Convolutional Neural Networks (CNN), such as You Only Look Once (YOLO). YOLOv8 provides multi-task capabilities, including simultaneous object detection and segmentation, known as instant segmentation through YOLOv8-Seg. The success of instant segmentation is influenced by the quality of the input image, because YOLOv8-Seg requires clear visual features such as texture, edges, and contrast. The image quality enhancement methods used are Discrete Wavelet Transform (DWT) to separate the image into frequency components based on local intensity variations in the image and Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance local contrast that has low contrast. Unsharp Masking is applied after CLAHE to sharpen edges and restore edge details that are reduced due to the DWT process. The combination of these methods produced a Peak Signal-to-Noise Ratio (PSNR) of 39.57 dB and a Structural Similarity Index Measure (SSIM) of 0.98, indicating excellent image quality. The instant segmentation results using YOLOv8-Seg achieved a precision of 80.16%, a recall of 98.47%, an Intersection over Union (IoU) of 79.17%, and an F1-Score of 88.38%, indicating that the model is capable of detecting nearly all colorectal polyp with a high degree of agreement of the segmented areas to the ground truth. The results of this study demonstrate that the combination of DWT, CLAHE, and UM methods can improve the quality of endoscopic images, so that the YOLOv8-Seg application is able to perform instant segmentation of colorectal polyps accurately and efficiently.
No other version available