Neural networks are one of the most powerful machine learning structures that have found many applications in various scientific and industrial fields in recent years due to the increasing computing power of processors. One of the advantages of these algorithms is that they make it possible to construct any arbitrary derivative function using only two hidden layers of neurons. Therefore, it is possible to learn any non-linear function using these structures.
In this project, we're going to investigate some issues related to neural networks with the TensorFlow library, using the Keras high-level interface. In this problem, we want to determine the animal species based on the image, using neural networks (MLPs).
The Oregon Wildlife Dataset is a wildlife image dataset. In this project, we will train a neural network with 4 classes of elk, raven, bald eagle and raccoon from this dataset.
Personal website: https://hesamasad.github.io