CXR-RefineDet: Single-Shot Refinement Neural Network for Chest X-Ray Radiograph Based on Multiple Lesions Detection

<div>Comparison of the proposed CXR-RefineDet and other mainstream object detectors. CXR-RefineDet at 512 resolution reaches 0.1686 mAP, surpassing all mainstream detectors, and its single image inference speed is nearly 2 times that of other detectors.</div>

Journal of Healthcare Engineering

fig1

Figure 1

Figure 1: CXR-RefineDet: Single-Shot Refinement Neural Network for Chest X-Ray Radiograph Based on Multiple Lesions Detection 