Improved Pneumonia Diagnosis of Radiological Images using Hybrid Loss with Conventional CNN

A bacterial infection is the root cause of pneumonia, a lung condition. Early diagnosis is crucial for the success of the therapeutic process. Chest X-ray images typically allow a qualified radiologist to make the diagnosis. The visual observation of the disease is one thing that can make a diagnosis challenging, and it can also be misdiagnosed for other concealed diseases in chest X-ray images. A diagnosis could be subjective due to other symptoms or disorders. The incorporation of computer-aided diagnostic tools improves the accuracy of pneumonia diagnosis and allow doctors to provide a second opinion before diagnosis. This research is focused on using a conventional simple convolutional neural network (CNN) with Hybrid Loss to create a processing model efficiently. The proposed model assists clinicians in identifying lesions on chest X-rays for accurate pneumonia diagnosis. This research showed that by incorporating the Hybrid Loss better results were obtained.