pyspacer 0.0.1

Spatial image analysis with caffe and pytorch backends.

PySpacer

This repository provide utilities to extract features from random point locations in images and then training classifiers over those features. It is used heavily with https://github.com/beijbom/coralnet.

Installation

Spacer has two installation modes. The full package requires the deep learning framework caffe to be installed. Since this can be a drag, caffe is only supported through docker.

Run including caffe using Docker

• Install docker on your system
• Build image: docker build -t "test:Dockerfile" .
• Run docker run -v /path/to/local/folder/:/workspace/models -it test3:Dockerfile

The -v /path/to/local/folder/:/workspace/models part will make sure the downloaded models are cached to your local disk (outside container), which makes rerunning stuff much faster.

This will run the default CMD command specified in the dockerfile (unit-test with coverage). If you want to enter the docker container do: docker run -it test3:Dockerfile bash.

Run without caffe support using virtulenv.

• Install virtualenv
• mkvirtualenv spacer --python /path/to/your/python3
• pip install -r requirements.txt
• python -m unittest

Code coverage

First generate report

coverage run --source=spacer --omit=spacer/tests/* -m unittest

Render simple with

coverage report -m

And to html with

coverage html

which renders html files to .htmlcov.