Open aux82716 opened 1 year ago
The fastest way is using the Tappas' gstreamer post processing implementation of that network.
בתאריך יום ו׳, 21 באוק׳ 2022, 16:06, מאת aux82716 @.***
:
Hi,
As a result of inference of the yolov5s_personface model I get 3 vectors of dimensions [1, 40, 40, 21], [1, 20, 20, 21], [1, 80, 80, 21]; what's the correct/fastest procedure to decode them in order to get a list of detections (such as [x_min, y_min, x_max, y_max, score, class])?
Thanks
— Reply to this email directly, view it on GitHub https://github.com/hailo-ai/hailo_model_zoo/issues/26, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADBIQYEIIH5LDW6NWUGU5R3WEKIMVANCNFSM6AAAAAARLE5XIY . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Can you share/point out the code to do that? I'm working in python with pyhailort
The Tappas pipeline called "Multi person multi camera detection" is using exactly this network to detect persons, and blur faces: https://github.com/hailo-ai/tappas/blob/master/apps/gstreamer/general/multi_person_multi_camera_tracking/multi_person_multi_camera_tracking.sh.
The post-processing of this is the same as using YOLOv5, with a minor change on the labels. This function is written in c++, here's the code: https://github.com/hailo-ai/tappas/blob/master/core/hailo/gstreamer/libs/postprocesses/detection/yolo_postprocess.cpp
we have an example notebook that binds this c++ function to python and run the full. Let me know if this is something that is of interest and I'll share through the support system, or privately.
Yes, can you share this notebook? Is there the possibility of using only the postprocessing detection script (in python) from the whole program? Thanks
Can you please send me an email to nadave@hailo.ai or open a request here: https://hailo.ai/contact-customer-support/
@nadaved1 I sent the mail to nadave@hailo.ai, reply plz. thanks
Hi,
As a result of inference of the yolov5s_personface model I get 3 vectors of dimensions [1, 40, 40, 21], [1, 20, 20, 21], [1, 80, 80, 21]; what's the correct/fastest procedure to decode them in order to get a list of detections (such as [x_min, y_min, x_max, y_max, score, class])?
Thanks