Enhanced Pneumothorax Detection in Chest Radiographs via a Deep Learning Framework DENSE Network Fusion with Yolov8

Pneumothorax (PTX, sometimes called collapsed lung, is a condition where air is trapped between the chest cavity and the lung. Human breathlessness is a potentially fatal condition known as pneumothorax. Every year, 1.88 million individuals pass away with pneumothorax. To diagnose PTX, chest radiographs, or CXRs, are utilized. Pneumothorax may be diagnosed by X-ray scans, which are a rapid, easy, and affordable procedure that is utilized in hospitals worldwide. The lack of skilled physicians and radiologists is the reason for the need for a technology that identifies PTX from X-ray images at advanced medical institutions. Professionals find it more challenging to handle manually, but it may be resolved automatically, simplifying the medical procedure. This paper proposes a framework for the automated and efficient diagnosis of pneumothorax from CXR utilizing a dense detection algorithm and pretrained YOLOv8. With a 98.72% accuracy rate, our system categorized CXRs as either PTX or healthy. The YOLOv8 detection method and the DensetNet-201 model fused increase classification performance while cutting down on training test time. There is a good chance that the suggested framework will be useful in helping medical professionals diagnose lung diseases.