Welcome to this workshop on Computer Vision and Convolutional Neural Networks. In a couple of steps we explain how computer vision techniques can be used to manipulate and analyse images and how convolutional neural networks (using KERAS) use transformations to analyse visual material.
The easiest way to install the requires libraries is through Anaconda. Make sure you have anaconda 5.2.0 installed for python 3.6.
You can download anaconda 5.2.0 here: https://repo.anaconda.com/archive/
Or you can make a python 3.6. environment in anaconda. This is described here: http://docs.anaconda.com/anaconda/faq/#how-do-i-get-anaconda-with-python-3-5-or-3-6
Clone this repository to your local machine using
git clone https://github.com/melvinwevers/CV_tutorial.git
Navigate to this directory and then input (replace new_environment with your preferred name for this environment, for example CV_course
conda create -n new_environment --file req.txt
If this does not work you can also create a new environment in the Anaconda GUI and install the following libraries by hand:
- pandas
- os
- Matplotlib
- skimage
- keras
- scipy
- glob
- tqdm
- sklearn
- face_recognition
Activate the environment in Anaconda or using your terminal (again replace new_environment with the name given to your environment)
conda activate new_environment
These instructions are for mac os. For windows please see these instructions: https://programwithus.com/learn-to-code/Pip-and-virtualenv-on-Windows/
Install virtualenv
pip3 install virtualenv
make directory for virtual environments (feel free to change this)
mkdir ~/virtualenvs
Make a virtual environment
virtualenv --system-site-packages -p python3 ~/virtualenvs/cv_course
activate environment
source ~/virtualenvs/cv_course/bin/activate
install iPython
pip3 install ipython
install Jupyter
pip3 install jupyter
install kernel
python3 -m ipykernel install --user --name cv_course --display-name "cv_course"
install libraries
pip3 install -r requirements.txt