International Journal of Electrical and Data Communication
  • Printed Journal
  • Refereed Journal
  • Peer Reviewed Journal

P-ISSN: 2708-3969, E-ISSN: 2708-3977

2024, Vol. 5, Issue 2, Part A


Artificial intelligence using YOLOv8 for the identification of elbow OCD in ultrasound images


Author(s): Devi Venkatesh Gowtham and Sweta S Munnoli

Abstract: When a pitcher uses their elbow over and over again, it might develop a condition known as baseball elbow. Baseball elbow diseases include osteochondritis dissecans (OCD), an intractable osteochondral lesion most often seen in middle and high school kids. Setting a period to cease playing baseball is the conservative therapy that may totally cure it if discovered in its early stages. Consultations are difficult to arrange due to the lack of discomfort during the early stages, and many patients experience a worsening of their illness. There is a scarcity of doctors who can diagnose baseball elbow, thus the number of modifications is several times each year, despite the fact that periodic medical checks are useful. For early identification and effective conservative treatment of elbow osteochondritis dissecans (OCD), ultrasonography (US) screening is necessary. Finding out how well YOLOv8, an AI model based on deep learning, can diagnose US photos of OCD or normal elbow-joint imaging is the main goal of the research. Methods: More than 2,430 photos were used. The YOLOv8 model was used for object identification and picture classification in order to identify OCD lesions or normal elbow joint images. End result: The confusion matrix values for the normal with OCD lesion binary categorization were: These metrics are as follows: F-measure = 0.9987, Accuracy = 0.998, Recall = 0.9975, and Precision = 1.000. Both the YOLOv8n and YOLOv8m models achieved mean average precision (mAP) values of 0.994 and 0.995, respectively, when comparing the trained model's bounding box detection to the true-label bounding box. In summary: Standard images of elbow joints while OCD lesions were used to train the YOLOv8 model for object identification and picture classification. The high degree of accuracy shown by both tasks suggests they might be valuable for baseball elbow screenings conducted as part of routine medical examinations.

Pages: 33-36 | Views: 1 | Downloads: 1

Download Full Article: Click Here

International Journal of Electrical and Data Communication
How to cite this article:
Devi Venkatesh Gowtham, Sweta S Munnoli. Artificial intelligence using YOLOv8 for the identification of elbow OCD in ultrasound images. Int J Electr Data Commun 2024;5(2):33-36.
International Journal of Electrical and Data Communication
Call for book chapter