In this project, we utilize a bone fracture datasets to train a neural network capable of accurately classifying bone breaks from X-ray images.
The dataset covers a range of bone fracture classes, such as avulsion fractures, comminuted fractures, fracture-dislocations, greenstick fractures, hairline fractures, impacted fractures, longitudinal fractures, oblique fractures, pathological fractures, and spiral fractures.
Automating the fracture classification process has the potential to enhance patient care and aid medical professionals in making well-informed decisions.
With Keras and TensorFlow I was able to construct a model that was able to detect bone fractures at 96% Accuracy
- Keras
- Tensorflow
- Python
- Multilayer Perceptron (76%)
- COnvolutional Neural Network (96%)
If you would like to visit the kaggle kernel this model was trained: https://www.kaggle.com/code/victor116/mlp-cnn-bone-break-classification-95-acc