<![CDATA[In this work, precision and recall measures are used to assess the performance of YOLOv7 and YOLOv8 models in identifying pulmonary carcinoma on a distinct collection of 700 photos. The necessity of early disease detection is increasing, thus choosing a reliable object detection model is essential. The goal of the research is to determine which model works best for this purpose, taking into account the unique difficulties that pulmonary cancer presents. The work makes a contribution to the field by showcasing the improvements made to YOLOv8 and underlining how well it detects both benign and malignant. YOLOv7 and YOLOv8 were used to independently train custom models using the pulmonary carcinoma dataset. The models' performance was measured using precision, recall, and mean average precision measures, which allowed for a comprehensive comparison examination. When it came to precision (58.2%), recall (61.2%), and mean average precision at both the 0.5:0.95 (33.3%) and 0.5 (53.3%) criteria, YOLOv8 outperformed YOLOv7. The 3.0% accuracy gain highlights YOLOv8's improved capabilities, especially in identifying small objects. YOLOv8's enhanced accuracy can be attributed to the optimisation of the detection process through its anchor-free design. According to this study, YOLOv8 is a more reliable model for pulmonary carcinoma identification than YOLOv7. The results indicate that YOLOv8 is the better option because of its higher recall, precision, and enhanced capacity to detect smaller objects—all of which are critical for early illness detection in medical imaging.]]>
Copyrights © 2024
