TRANSFORMER DECODER FOR CHEST X-RAY IMAGE CAPTIONING USING DEEP FEATURE EXTRACTION

Chest xray are widely used in hospital to help doctors diagnose lung problems. Since the outbreack of covid19, especially during second wave and winter season, it has become even more important to quickly detect the disease. To help doctors and reduce their workload, we use Deep Learning to automatically analyze chest X-ray .This study, propose a method can look at a chest X ray image and automatically generate a medical report. First, we use a model called Vision Transformer (ViT) to understand overall features of image. use another model called CheXNet, which is good at identifying chest related diseases, to extract detailed medical features.These features are combined and sent to a Transformer decoder,which creates a meaningful text description of what is seen in the image. This helps doctors by giving them a quick, accurate summary of the patients condition, making the diagnosis process faster and more reliable.