TO ASSESS A BOUNDARY DETECTION ALGORITHM

(1) Run the boundary detection algorithm in one of the images from '/benchmark/test/' and save the result as '.bmp' inside '/benchmark/results/X/', where X is short name (no spaces) that reminds you of the method and parameters used. Important: your resulting '.bmp' HAS to be named exactly as the test image.

(2) Compile the main code run.cpp: type 'g++ -o run run.cpp'.

(3) Run a test: type 'run <path_to_resulting_images> <file_with_indices_to_process>'. 
E.g. 'run /home/user/match/benchmark/results/alg1 iids_animals.txt'         
     'run /home/user/match/benchmark/results/alg1 iids_humans.txt'.
     'run /home/user/match/benchmark/results/alg1 iids_manmadeobjects.txt'             
     'run /home/user/match/benchmark/results/alg1 iids_natural.txt'

(4) Read the results (Recall and Precision rates) displayed on the screen and saved as '<index_to_image>_rp.txt' in your algorithm's folder.

Note: If an error occurs (e.g. inexistant path, wrong image index), type 'exit' to go back to the Linux prompt.