This repo contains a basic python course for the image and signal processing group at IPL. It covers basic syntax and some scientific python packages like numpy, matplotlib and sklearn. It focuses in image classification.
The goal will be to make some predictions for the dogs vs cats classification problem kaggle competition
- Laptop
- Download in advance:
- Anaconda for python 3.5 https://www.continuum.io/downloads
- Train images: https://www.dropbox.com/s/8lbkqktfofzjraj/train.zip?raw=1
- clone/download this repo.
The workshop outline is the following:
The workshop is divided in two parts dessigned to be covered in two days (~3-4h each day):
- Preliminars:
- What's python?
- Why python?
- Installing python
jupyternotebook basics
- Python language (
basic_syntax.ipynb):- Basic syntax and core structures (loops, functions, lists, ifs, strings...classes, dicts)
- file operations (list directory, r/w files, get data from web,...)
- Example dogs vs cats.
numpy(numpy.ipynb):- basic matrices
- read image as matrix
- basics on matrix transformations.
scikit-learn(machine_learning.ipynb):- predict label on image
- hyper-parameter selection via cross-validation
- show prediction metrics (accuracy, ROC, confussion matrix).
- Deep learning (
deep_learning.ipynb):- Use pretrained CNN with
keras.
- Use pretrained CNN with
- Conclussions:
- my way of working.
- Where to go to learn more.