search

apennisi / fastestpedestriandetectorinthewest

This is my version of the Fastest Pedestrian Detector in the West
GNU General Public License v3.0
24 stars 23 forks source link

Two different results are obtained for the same image when using C++ version and MATLAB version #5

Open TAO270 opened 6 years ago

TAO270 commented 6 years ago

A new pedestrian detector was obtained by training using Piotr's toolbox, and then it was converted into .xml file via the function savetoxml(). However, when I used this .xml file to detect pedestrian in a picture including one person, I could not get the right result. So, I performed the same operation on the Matlab toolbox for the same picture, and　the position of the person was correctly marked. C++ version. 33 Matlab version 11

TAO270 commented 6 years ago

In addition, the detection was performed in Linux, and the detector was trained with some gray images if possible, I was wondering why there is such a big difference. Thank you in advance.

TAO270 commented 6 years ago

Here is my own detector in format TXT. inria_detector_gray.txt