Published Paper Details:

Pneumonia is a respiratory disease caused by bacteria or viruses.it affects many people, especially in developing and developing countries, where pollution levels are high, unhealthy living conditions and congestion are common along with inadequate medical infrastructure. Pneumonia causes pleural leakage, a condition in which fluid leaks out fill the lungs, causing difficulty breathing. Early diagnosis of pneumonia is important to confirm therapeutic treatment and increasing survival rates. Chest X-ray imaging is the most common method used to diagnose pneumonia. However, chest X-ray examination is a challenging work and prone to tangible flexibility. Pneumonia remains a threat to human health; Coronavirus 2019 (COVID-19) which started in late 2019 has had a huge impact on the world. It continues in many lands and has caused tremendous losses in human life and property. In this paper, we present the DeepConv-DilatedNet-based method of diagnosis and locating pneumonia in chest X-ray images (CXR). The two-phase Faster R-CNN detector was adopted as a network infrastructure. Feature Pyramid Network (FPN) is integrated into an expanded neural network of bottleneck extensions to deepen features to preserve in-depth feature and location information. In the case of DeepConv-DilatedNet, the deconvolution network is used to restore high-level feature maps to their original size, and targeted information is stored. DeepConv-DilatedNet, on the other hand, uses the most popular full-featured layouts and shares some calculations throughout the image. Then, Soft-NMS is used to check boxes and verify sample quality. Also, K-Means ++ is used to produce work boxes to improve the precision of spatial processing. The algorithm obtained 39.23% Mean Average Precision (mAP) from the X-ray image database from the Radiological Society of North America (RSNA) and obtained 38.02% Mean Average Precision (mAP) from the ChestX-ray14 database, surpassing other information algorithms. Therefore, in this paper, an advanced algorithm that can provide doctors with information on the location of pneumonia lesions is proposed.