YOLO v3 and YOLO v4 Comparison Video With Deep SORT

preview_player
Показать описание
YOLO v3 and YOLO v4 Comparison Video With Deep SORT

The white boxes are Deep SORT trackers and the blue boxes are YOLO v4 detections. Each white box has a tracking ID at the top and each blue box has a YOLO detection confidence score at the bottom.

Same settings used for both:
Confidence threshold = 0.5
IoU threshold = 0.5
Image size: 416 x 416

Real-time FPS:
YOLO v3: ~4.3fps
YOLO v4: ~10.6fps

YOLO v4 performs much faster and appears to be more stable than YOLO v3. All tests were done using an Nvidia GTX 1070 8gb GPU and an i7-8700k CPU.

#machinelearning #computervision #yolo #deepsort #leonlok

STAY UPDATED:

MY GEAR:
These are affiliate links for the gear that I use (I receive a kickback).

ABOUT ME:
I'm Leon and I'm a data scientist based in the UK. I create videos about data science, and I try to share my knowledge and life experience to help others grow 🙂

GET IN TOUCH:
Рекомендации по теме
Комментарии
Автор

The problem of software isn't coding but the hardware. No viable, it will not work

thegamechanger