Computer vision in the new era of Artificial Intelligence and Deep Learning

Visión por computador en la nueva era de la Inteligencia Artificial y el Deep Learning

This repository contains a collection of Jupyter Notebooks as well as other resources (e.g. presentations, links, ...) that are going to be used during the "Second quarter university extension course" that the University of Oviedo is going to teach (online). For more information about this course, you can check the following URL

Work in progress 🚧

Objectives

✔️ To know the current context of computer vision in the new era of artificial intelligence and Deep Learning

✔️ To know the main tools and strategies for problem solving using different data sources

✔️ Knowing how to handle computer tools and specific software for digital image processing

✔️ Understanding how basic digital image processing and analysis methods and techniques work

Content

✔️ Introduction to OpenCV and Python

✔️ Basic image treatment

✔️ Image processing

✔️ Machine learning (Scikit-learn)

✔️Deep learning (Tensorflow, Keras, PyTorch)

The target audience is broad and includes

People who have experience in computer science (maybe to graduate level) but who do not know about OpenCV
People who are studying other subjects and want to play with computer vision

Introduction

Presentation

Notebooks

Colab utilities and tricks for computer vision

Lesson	Estimated time needed	Source Code	Colab	Presentation
Colaboratory introduction	10 min	Open	Open	Presentation
Collection of features, utilities and tricks on Colab and/or Python	30 min	Open	Open	Presentation
Take image from webcam as numpy array in Colab	30 min	Open	Open	Presentation
Take video from webcam in Colab	30 min	Open	Open	Presentation
Explore, execute and see the output of external Python scripts	30 min	Open	Open	Presentation
Create and show multiple images in the same figure with matplotlib	15 min	Open	Open	Presentation
Install Colab utilities (https://github.com/albertofernandezvillan/colaboratory-utils)	30 min	Open	Open	Presentation

Introduction to Python and NumPy

Lesson	Estimated time needed	Source Code	Colab	Presentation
Introduction to Python	30 min	Open	Open	Presentation
Introduction to Numpy	30 min	Open	Open	Presentation

OpenCV notebooks for computer vision

OpenCV basics

Lesson	Estimated time needed	Source Code	Colab	Presentation
Basic image treatment (color and grayscale images) in OpenCV	30 min	Open	Open	Presentation
BGR color format in OpenCV	30 min	Open	Open	Presentation
Drawing basic figures (points, lines, poligonal curves,...) in OpenCV	30 min	Open	Open	Presentation
Drawing text and symbols in OpenCV	10 min	Open	Open	Presentation

Image processing with OpenCV

Lesson	Estimated time needed	Source Code	Colab	Presentation
Geometric image transformations in OpenCV	15 min	Open	Open	Presentation
OpenCV sliders for image processing	20 min	Open	Open	Presentation
Visual interfaces (with buttons and sliders) for image processing with OpenCV	20 min	Open	Open	Presentation
Computational photography module in OpenCV	20 min	Open	Open	Presentation
Inpainting with OpenCV	20 min	Open	Open	Presentation
K-means clustering for color quantization with OpenCV	20 min	Open	Open	Presentation
References for main image processing techniques and algorithms in OpenCV	50 min	Open	Open	Presentation
Some packages for face processing in Python	20 min	Open	Open	Presentation

Configure OpenCV with GPU on Colab and benchmarking inference speed

Lesson	Estimated time needed	Source Code	Colab	Presentation
Configure OpenCV with GPU con Colab	20 min	Open	Open	Presentation
Benchmarking GPUS vs CPU with OpenCV in Colab (human pose estimation)	20 min	Open	Open	Presentation
Benchmarking GPUS vs CPU with OpenCV in Colab (YOLO V4)	20 min	Open	Open	Presentation

Machine learning for computer vision

Introduction to scikit-learn for classification, regression and clustering

Lesson	Estimated time needed	Source Code	Colab	Presentation
Scikit-learn introduction for classification	40 min	Open	Open	Presentation
Scikit-learn introduction for regression	30 min	Open	Open	Presentation
Scikit-learn introduction for clustering (color quantization)	20 min	Open	Open	Presentation
Closed eyes detection	30 min	Open	Open	Presentation

Introduction to pandas

Lesson	Estimated time needed	Source Code	Colab	Presentation
Pandas introduction	30 min	Open	Open	Presentation
Minimal example using both scikit-learn and pandas for classification	30 min	Open	Open	Presentation

Introduction to metrics in scikit-learn

Lesson	Estimated time needed	Source Code	Colab	Presentation
Metrics for classification	30 min	Open	Open	Presentation

TensorFlow and Keras

Lesson	Estimated time needed	Source Code	Colab	Presentation
Simple MNIST convnet	15 min	Open	Open	Presentation
Simple MNIST convnet (2)	15 min	Open	Open	Presentation
Using a pre-trained model for inference using Keras Applications	15 min	Open	Open	Presentation
Set up Kaggle API in Colab	5 min	Open	Open	Presentation
Using Keras Applications for feature extraction in a classification problem with scikit-learn	30 min	Open	Open	Presentation
Using Keras Applications for feature extraction in a clustering problem with scikit-learn	25 min	Open	Open	Presentation
Image and dataset augmentation	20 min	Open	Open	Presentation
Transfer learning using Keras Applications	35 min	Open	Open	Presentation

PyTorch

Lesson	Estimated time needed	Source Code	Colab	Presentation
Tensors	30 min	Open	Open	Presentation
Learning	60 min	Open	Open	Presentation
Neural networks	40 min	Open	Open	Presentation
Transfer learning	40 min	Open	Open	Presentation

Lesson	Estimated time needed	Source Code	Colab
Minimal example showing how to create and deploy a Flask app for computer vision	20 min	Open	Open

Lesson	Estimated time needed	Datasets	Models	Presentation
Datasets and models	50 min	Open	Open	Presentation

Other resources

Additional packages and libraries for computer vision and image processing in Python

Introduction to Pillow: Python Imaging Library (PIL Fork)

Lesson	Estimated time needed	Source Code	Colab	Presentation
Pillow: Python Imaging Library (PIL Fork)	20 min	Open	Open	Presentation

Scikit-image (Image processing in Python)

Repository	description
scikit-image	scikit-image: Image processing in Python
skimage-tutorials	scikit-image tutorials

Stéfan van der Walt, Johannes L. Schönberger, Juan Nunez-Iglesias, François Boulogne, Joshua D. Warner, Neil Yager, Emmanuelle Gouillart, Tony Yu, and the scikit-image contributors. scikit-image: Image processing in Python. PeerJ 2:e453 (2014) https://doi.org/10.7717/peerj.453

Mahotas: computer vision in Python (Image processing and computer vision in Python)

Repository	description
mahotas	mahotas: Python Computer Vision Library
mahotas-demos	mahotas tutorials and demos

Luis Pedro Coelho Mahotas: Open source software for scriptable computer vision in Journal of Open Research Software, vol 1, 2013. [DOI]

Additional packages and libraries for machine learning in Python

Dlib (C++ toolkit (with python API) containing machine learning algorithms and tools)

Repository	description
dlib	dlib library
Examples: Python	mini-tutorials for using dlib from Python (face detection, tracking, recognition or alignment)

KING, Davis E. Dlib-ml: A machine learning toolkit. The Journal of Machine Learning Research, 2009, vol. 10, p. 1755-1758.