PyImageSearch / imutils

A series of convenience functions to make basic image processing operations such as translation, rotation, resizing, skeletonization, and displaying Matplotlib images easier with OpenCV and Python.
MIT License
4.51k stars 1.03k forks source link

Compatibility of vectorized non_max suppression code with Ultralytics Yolov3 non max suppression code? #239

Open Mechazo11 opened 3 years ago

Mechazo11 commented 3 years ago

Dear @jrosebr1,

Thank you for making these awesome object detection codes available publicly online.

I am working on an people detection project and am using Ultralytics YoloV3 as the base code. In his general.py file, the author uses a form of NMS which takes a Pytorch tensor containing all the bounding box proposals as inputs. The code can be found here https://github.com/ultralytics/yolov3/blob/master/utils/general.py

I was wondering is your implementation of nms compatible with Ultralytic's NMS code?

With best regards,

jrosebr1 commented 3 years ago

Sorry, I haven’t tested with that particular implementation. Give it a try and see. Experimentation is the best way to learn.

On Jan 29, 2021, at 10:15 AM, AZMYIN MD KAMAL notifications@github.com wrote:

 Dear @jrosebr1,

Thank you for making these awesome object detection codes available publicly online.

I am working on an people detection project and am using Ultralytics YoloV3 as the base code. In his general.py file, the author uses a form of NMS which takes a Pytorch tensor containing all the bounding box proposals as inputs. The code can be found here https://github.com/ultralytics/yolov3/blob/master/utils/general.py

I was wondering is your implementation of nms compatible with Ultralytic's NMS code?

With best regards,

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.