Segmentation and Classification of Acute Respiratory Distress Syndrome (ARDS) using Deep Learning

Acute Respiratory Distress Syndrome (ARDS) is a serious respiratory syndrome with incidence rates varying based on underlying factors like sepsis, covid and pneumonia. It can impact people of all ages, often necessitating mechanical ventilation. Mortality rates are more and influenced by factors like age, existing health conditions, and the triggering event's nature. Prompt intervention and addressing the root cause are pivotal. The manual segmentation of the lungs leads to workload for the workers. So our aim is to provide automatic lung segmentation with a high amount of accuracy and precision. Here the convolutional neural network was used to train the deep learning model using 257 sample affected images. Among these images 80% are used for training and the remaining 20% is used for testing the model. The trained deep learning model is evaluated using intersection over union (IOU). On the testing data set, the automatic segmentation gives the accuracy of 99% and the IOU index of 0.99. The trained model gave fast and precise results. The model can automatically segment the lungs with a high amount of precision.