Sequential CNN Model for Hemorrhage Prediction using Brain CT Images

Haemorrhage diseases, sometimes called bleeding, is the flow of blood out of blood vessels as a result of a break in the wall of the vein. With respect to the source and location of the bleeds, this can happen either internally or externally and range in severity from minor to serious. In order to predict the existence of bleeding, a convolutional neural network (CNN) model was trained on a dataset consisting of 6094 CT scan pictures in this study. An additional test set consisting of 678 photos was used to assess the model. The model attained a training accuracy of 90.36% and a training loss of 0.2833 after 50 epochs of training with the RMSprop optimizer. The model achieved 91.63% validation accuracy and a validation loss of 0.2448 on the validation set. These findings show that the CNN model is very accurate and has a low loss in identifying haemorrhage in brain CT scan images. The model's strong performance was aided by the efficient parameter optimisation made possible by the RMSprop optimizer.