• Dominique MICHEL
• Léa Masselles
• Louis Guo
• Kenny Fung

• Cmake 3.17+ for building purposes
• Boost : program_options
• OpenCv: image, video, gui
• Cuda Toolkit: nvcc

VERY IMPORTANT, the "kmeans.database" file must be in the working directory when executing the program

See the python code You can create the lbp matrix of histograms from this binary using the -l or --gen-lbp option

$ ./BarcodeDetector -l input_file.jpg

There should be a input_file.jpg.txt This file can be import in python using OpenCv as following:

import cv2
fs = cv2.FileStorage("input_file.jpg.txt", cv2.FILE_STORAGE_READ)
fs.getFirstTopLevelNode().mat()

If you wish to use different than default classes (3, 12 ,15), you must edit the following source file function :

src/my_opencv/wrapper.cc::isCorrectClass

You can run a benhmarking binary by running the following command in the build directory:

$ make bench
$ ./bench/bench