search

cyrusbehr / YOLOv8-TensorRT-CPP

YOLOv8 TensorRT C++ Implementation
MIT License
568 stars 70 forks source link

Getting inference data from these results #43

Closed kazmifactor closed 9 months ago

kazmifactor commented 9 months ago

Is there any way to get the inference data from these results. Like coordinates of the bbox. Or keypoints in pose. In live video

cyrusbehr commented 9 months ago

Hi @kazmifactor

Please spend some time looking through the code. You'll notice that all the API methods return std::vector<Object> which contains the information you are after: https://github.com/cyrusbehr/YOLOv8-TensorRT-CPP/blob/afe5a445a64869fb0c5f690ce0d58dc7ea625a41/src/yolov8.h#L75-L81

kazmifactor commented 9 months ago

Thank you for your reply.

I am not able to decode it and get the data that is required for my application. i am very new to the lniux and c++ framework.

if you could help me with making a separate code that runs a video file and sends each frame separately, and get the results like bbox or keypoints either through python or c++.

this would help me or any other person who will use this in future. Thank you in advance.

cyrusbehr commented 9 months ago

My code already works with a video input, decodes the outputs, and annotates the video. It already does everything you are asking for. I can't help you any more than that. The project is already fairly simple. You need to spend some time learning C++ and try running the project. I cannot help you with that.