Enhancing Haemorrhage Detection in Head CT Scans Using Deep Learning
Aditya Kumar, Leema Nelson, Sudhakar Kumar
Abstract
This work investigates the application of deep learning for haemorrhage detection in head CT scans. The aim of this work is to develop a robust model for accurate detection, even with limited data. The Sequntional CNN model is utilised to detect the haemorrhage CT scans, consisting of convolutional layers to extract features, dense layers for classification, pooling layers for dimensionality reduction, and dropout layers for regularisation. This model uses the data augmentation technique to enhance the training data through transformations like rescaling, shearing, and rotation. This model mitigates overfitting and enhances the overall robustness of the model. Moreover, this model provides significant improvement in accuracy, reaching 94.99%. This performance highlights data augmentation, enhancing the generalisation capacity of the model. The developed model outperforms other deep learning models used in similar applications. This work highlights the adaptability of deep learning in medical image analysis and its potential to enhance healthcare applications, with implications for improved patient outcomes.