A problem when using Smart Classroom Demo

[Hiiragi_Iori]

Today I wanted to run Smart Classroom Demo while a problem occured. I ran the demo as follows:

./attendance_system_recognition -m_fd /home/hiiragiiori/attendance_system/model/face-detection-adas-0001.xml -m_reid /home/hiiragiiori/model/face-reidentification-retail-0095.xml -fg /home/hiiragiiori/faces_gallery.json -i cam

Then there was a warning says "Face recognition models are disabled!", and the labels couldn't be read either.

I re-downloaded the models with "downloader.py" and the same warning occured. But it went well about a couple months ago. 

Please tell me how to solve this problem. Thanks a lot!

[Shubha_R_Intel]

Dear Hiiragi Iori,

Are you sure you passed in all the command-line args correctly ? Please review the following document:

http://docs.openvinotoolkit.org/latest/_demos_smart_classroom_demo_README.html

In Smart Classroom Demo main.cpp if you search "Face recognition models are disabled!" it lands in model paths not being passed in correctly.  The error is here:

 if (!reid_config.enabled) {
            slog::warn << "Face recognition models are disabled!"  << slog::endl;
        } else if (!face_gallery.size()) {
            slog::warn << "Face reid gallery is empty!"  << slog::endl;
        } else {
            slog::info << "Face reid gallery size: " << face_gallery.size() << slog::endl;
        }

But if you backtrack and search reid_config and so on, you will see that it has to do with model paths which are passed in wrongly.

Let me know if it helps,

Thanks,

Shubha

[Hiiragi_Iori]

Dear Shubha,

I did pass in all the model paths correctly.

Actually it's working well on my own system, but doesn't work well on the Intel FPGA server. I'll check it again. Maybe there's some other problems.

Thanks a lot!

[sun__adam]

Dear all 

We face the same problem "Face recognition models are disabled!"

OpenVINO version 2019.R3.1

OS: Ubuntu18.04

execute the Smartclassroom and we make sure we pass the correct path about face-reidentification-retail-0095

is that any update for this problem??

thanks a lot

[Hiiragi_Iori]

Dear sun, adam

I've solved this problem some days ago.

If you want to do face reidentification, you should load at least these three models:

1、face-detection-adas-0001

2、face-reidentification-retail-0095

3、landmarks-regression-retail-0009

Then you will get the results.