Genetic Edge Detector

The Genetic Edge Detector is an application which implements two genetic algorithms for the Edge Detection. These implementations have been created on the basis of the articles:

• An edge detection technique using genetic algorithm-based optimization, written by Suchendra M. Bhandarkar, Yiqing Zhang and Walter D. Potter (available at the researchgate portal);
• Optimal edge detection filter using genetic algorithm, written by R. M. Naife and H. H. Abass (available on the Web).

Requirements

You need to have a python 3.4+ distribution installed (tested on the 3.5.2 version). If you want to install dependencies defined in the requirements.txt file, you also need to have a proper version of pip installed (tested on 8.1.2 and 9.0.1 versions for the python 3.5).

Installation and Commissioning

Configuration files for both algorithms are included in the resources/config directory. First of all, please specify required values (i.a. paths of input and output files and algorithms' parameters) in the appropriate configuration file. Then, open the command prompt in the main directory of the application and type:

pip3 install -r requirements.txt
python3 run.py 

Please note that the algorithm to be used may be changed simply by editing the run.py script.

Running the Canny edge detector

There is an implementation of the Canny edge detector provided by the scikit-image library. The configuration file for the script performing Canny edge detection is placed in the resources/config directory. In order to run the script, open the command prompt in the main directory of the application and type:

python3 canny/canny_detection.py

Note that it is necessary to have the dependencies installed already.

Running the performance evaluators

This repository is equipped also with the scripts responsible for evaluation of the algorithms' performance (based both on produced images and log files). The configuration file for evaluation scripts is included in the resources/config directory. Firstly, please specify paths to input files which will be evaluated. Afterwards, open the command prompt in the main directory of the application and type:

python3 evaluation/image_evaluation.py

to run the script evaluating output images, or:

python3 evaluation/logs_evaluation.py

to run the script evaluating log files produced by the algorithms. It is also necessary to install the dependencies before.