Page tree
Skip to end of metadata
Go to start of metadata

Introduction

SLDC [Mormont et al., Benelearn, 2016] is a generic framework for accelerating development and execution of multi-gigapixel images analysis workflows., with a focus on problems of object detection and classification. It enables developers to implement arbitrarily complex workflows without caring about problem-independant concerns such as large image handling and parallelization. Developers can plug their segmentation and classification algorithms into SLDC which will then provide them with the detected objects and their classification labels. 

Installation procedure

First install the Cytomine-Python-Client and Cytomine-Python-DataMining, see instructions here.

cd $HOME/Cytomine (or your Cytomine base directory)
source activate cytomine
git clone https://github.com/waliens/sldc.git
cd sldc
python setup.py build
python setup.py install

To check it works:

python -c "import sldc"

To install the Cytomine binding so that SLDC can access Cytomine image data: 

cd $HOME/Cytomine/Cytomine-python-datamining/cytomine-datamining/algorithms/sldc/
python setup.py build
python setup.py install

To check it works:

python -c "import cytomine_sldc"

Main algorithm

The main algorithm implemented in SLDC works as follows: 

  1. Split the image into tiles
  2. Segment each tile using a segmentation procedure S defined by the developer
  3. For each tile, extract polygons representing the detected objects
  4. Merge all polygons overlapping several tiles
  5. Dispatch each polygon to its most appropriate classifier using a dispatching procedure D defined by the developer
  6. Classify each polygon using some classification procedures Si and to a label and its probability per polygon

More details in Chapter 3.1 of [Mormont R., Master thesis, 2016].

Documentation and tutorials

Documentation, tutorials and examples about SLDC:

  • No labels