/myGAsaliency

Primary LanguageC++

myGAsaliency

This project is about optimizing the weights of a computational framework that detects salient objects in camera frames.

My framework

  • The algorithm used for optimization is a Genetic Algorithm.
  • The weights [w_1, w_2, w_3] are represented in binary strings.
  • The objective function used is the Euclidean Distance between the color images and their equivalent groundtruth in the SED1 and SED2 databases.
  • The parents are selected using Tournament Selection.
  • A 2-point Crossover is implemented.
  • Flip-bit Mutation is implemented.

Implementation Details

The framework is implemented in C++. The following libraries are used:

  • C++ Standard
  • xtensor
  • OpenCV

You can compile the code with:

g++ -I "/home/usr/ProjectDirectory/xtl/include" -o myGAsaliency my_GA_saliency_framework.cpp `pkg-config opencv --cflags --libs`