Building a detection model using a Convolutional Neural Network in Tensorflow & Keras. Used a brain MRI images dataset founded on Kaggle.
The dataset contains 2 folders: The folder yes contains 155 Brain MRI Images that are tumorous and the folder no contains 98 Brain MRI Images that are non-tumorous.
There wasn't enough examples to train the neural network. Also, data augmentation was useful in taclking the data imbalance issue in the data.
Before data augmentation, the dataset consisted of: 155 positive and 98 negative examples, resulting in 253 example images.
After data augmentation, now the dataset consists of: 1085 positive and 980 examples, resulting in 2065 example images.
Note: these 2065 examples contains also the 253 original images. They are found in folder named 'augmented data'.
For every image, the following preprocessing steps were applied:
- Crop the part of the image that contains only the brain.
- Resize the image to have a shape of (240, 240, 3). So, all images should have the same shape to feed it as an input to the neural network.
- Apply normalization
- 70% of the data for training.
- 15% of the data for validation.
- 15% of the data for testing.
Now, the model detects brain tumor with:
-
83.6% accuracy on the test set.
-
85.87% f1 score on the test set.