Since we're planning to implement a repeated random cross-validation, we'll need to be able to run our C code and MATLAB code side by side, with the same dataset partitioning, so that we can have comparable results. To do this we need to modify the MATLAB code to consume the image database in the same way as the C code. Then we can have our cross validation script, on each iteration, prepare the training/test sets, run the MATLAB code, run the C code, and compare the results.
Since we're planning to implement a repeated random cross-validation, we'll need to be able to run our C code and MATLAB code side by side, with the same dataset partitioning, so that we can have comparable results. To do this we need to modify the MATLAB code to consume the image database in the same way as the C code. Then we can have our cross validation script, on each iteration, prepare the training/test sets, run the MATLAB code, run the C code, and compare the results.