An AI based identification of non-complying or Pneumonia-infected individuals by classifying CT scan images. It can be used as preliminary test for people showing symptoms to cut the strain on existing medical facilities.
We've implemented transfer learning using pre trained models like InceptionV3 and VG19. They speed up the time it takes to develop and train a model by reusing these pieces or modules of already developed models. This helps speed up the model training process and accelerate results.
This image recognition model consists of two parts:
- Feature extraction part with a convolutional neural network
- Classification part with fully-connected and softmax layers
In transfer learning, we build a new model to classify the original dataset, where we reuse the feature extraction part and re-train the classification part with your dataset.
The VGG16 model was developed by the Visual Graphics Group (VGG) at Oxford. By default, the model expects color input images to be rescaled to the size of 224×224. The model is specifically trained for a more than a million type of images
Similar to VGG16, the main difference between the “VGG-19 Neural Network” and the “VGG-16 Neural Network” is that, this type of network is 19 layers deep and that type of network was 16 layers deep respectively.
Metrics | InceptionV3 | VGG16 | VGG19 | |||
---|---|---|---|---|---|---|
0 | 1 | 0 | 1 | 0 | 1 | |
Precision | 0.41 | 0.63 | 0.37 | 0.62 | 0.37 | 0.62 |
Recall | 0.23 | 0.80 | 0.20 | 0.79 | 0.19 | 0.81 |
F1-score | 0.30 | 0.71 | 0.26 | 0.70 | 0.25 | 0.70 |