In this example, we use the pre-trained ResNet50 and VGG16 model, which is pretrained on the ImageNet dataset. The implementation is in TensorFlow-Keras.
- We download the weights from "https://github.com/fchollet/deep-learning-models/releases" (resnet50_weights_tf_dim_ordering_tf_kernels_notop_updated.h5).
- We set the paths to the train, validation and test files, with respect to our local system path.
- Using the pretrained Resnet50 model, we fine-tune the model according to our own dataset, and then extract the features from the entire dataset.
- We save the features as a .Mat file.
- Tensorflow
- Numpy
- Scipy.io
Happy coding!!