Skripsi

OPTIMASI PRECISION PADA LESI HISTOPATOLOGI HATI MENGGUNAKAN YOLOV11

---

Rapid and reliable pathological evaluation is made possible in large part by the automatic detection of liver histological abnormalities. Using YOLOv11 on the Roboflow Liver Disease public dataset, this work aims to maximize precision in the detection of four lesions: ballooning, fibrosis, inflammation, and steatosis. After pre-processing the data with a 640×640 resize, the model was trained using the following configurations: epochs equals 10, batch equals 4, imgsz equals 640, lr0 equals 0.01, and the operating points are evaluated using parameter configuration adjustments, especially on IoU-NMS and confidence threshold. On the validation set, the model obtained mAP@50 equals 54.7%, Precision equals 80.1%, and Recall equals 28.3% at confidence equals 0.45 and IoU equals 0.60. With a lower recall as a result, these findings validate the effectiveness of the precision-oriented approach (reducing false positives). Class-wise analysis reveals that ballooning and steatosis can be detected more easily, but fibrosis and inflammation are still difficult to identify because of their tiny size and poor contrast. In addition to outlining areas for improvement (tiling/high resolution, TTA, hard negatives/positives, and class-wise thresholding) for future study aiming for higher recall without compromising precision, this work presents a repeatable precision-oriented pipeline and operational points. Keywords: YOLOv11, Object Detection, Liver histopathology, swelling, fibrosis, inflammation, steatosis, precision, confidence threshold, IoU-NMS, Roboflow.

---

Availability

Detail Information

Series Title

-

Call Number

T1962742026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xx, 96 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.370 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Computer Vision
Prodi Teknik Informatika

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment