Evaluation of YOLOv8 in Lung Lesion Segmentation on CT Images

Lung cancer remains the leading cause of cancer death worldwide, highlighting the urgent need for advanced diagnostic tools. This paper evaluates the performance of You Only Look Once version 8 (YOLOv8), an advanced object detection model, for segmenting lung lesions in Computed Tomography (CT) images. By leveraging the real-time detection capabilities of YOLOv8, we aim to improve the accuracy of lung lesion identification, a key factor in early diagnosis and effective treatment planning. The model was trained and tested on a comprehensive dataset of CT scans, focusing on optimizing both segmentation accuracy and lesion location. The results show that YOLOv8 significantly enhances the efficiency of lesion detection, demonstrating significant improvements in sensitivity and specificity when compared to traditional methods. This proposal highlights the potential of YOLOv8 in medical imaging and its role in advancing automated lung cancer screening systems. The model’s ability to streamline lesion identification offers promising advances for earlier diagnosis, better treatment outcomes, and overall improvements in clinical decision-making processes.