This is the implementation of 'Sparse Representation based Open-set Recognition'
-
In 'train_together.m'. Generate your training samples and testing samples, Proceed training and save the training detail
-
In 'objsect_src_evt.m'. Generate the tail distribution (GPD) of matched and sum of non-mathced reconstruction errors using the In this code, we also do testing using SRC. The testing result will be saved. The tail distribution of matched and sum of non-mathced will be saved.
-
In 'object_src_fmeas.m'
Calculate the F-measure and Accuracy using 'object_src_fmeas.m' -
Make sure to specify all the data-related parameter based on your data. (such as tail size, weights, thresholds)
** We also include one sample in the code. You can directly calculating the F-measure and Accuracy by running 'object_src_fmeas.m'
** All the code is writen in Ubuntu 14.04.
** If you want to run the demo, feel free to contact me if you need my data via email : [email protected]