Image Recognition with Deep Learning

BY: Nedal Ahmad

Aquincum Institute of Technology "AIT-Budapest", Fall semester 2021

Subject: Deep Learning course

Instructor:Professor Bálint Gyires-Tóth

Abstract:

Deep learning has significantly improved image recognition systems which can be applied in every aspect of our lives. In this project, I will build a Deep Neural Network (DNN) to recognize certain objects in some photo using CIFAR-10 dataset to predict one of ten classes of objects. adjust state of the art DNN for recognition of other new-objects but we don't need to do training again.

Requirments:

Python 3.6.6

Microsoft Visual C++ 2015 Redistributable update 3

PyCharm v. 2018.2

CIFAR_10 for windows x64

import Libraries:

1. numpy

2. pip>=9.0.0

3. pandas

4. matplotlib

5. Keras==2.1.6

6. h5py

7. pillow

8. scikit-learn

9. scipy

10. tensorboard==1.8.0

11. tensorflow==1.8.0

12. google-api-python-client==1.6.7

13. joblib

Dataset:

CIFAR10 http://www.cs.toronto.edu/~kriz/cifar.html