Automated Diagnosis of Spine Pathologies using Deep Convolutional Neural Networks (CNNs)

Diagnosis of spine diseases can be a problem in terms of inaccuracy and speed with manual approaches. This work is going to introduce the classification of spine problems-such as Scoliosis and Spondylolisthesis-by using an automated technique that includes deep Convolutional Neural Networks working over medical images. This approach will feed images through augmentation, normalization, and scale all images evenly before extracting as much relevant information as possible from the image using CNN. It has achieved this model by training and evaluating on labelled datasets to perform well in terms of accuracy. Future work will involve integrating the models in a clinical context to extend the set of conditions that can be detected.