AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.79k stars 7.97k forks source link

data augmentation using rotation #4776

Open th-yong opened 4 years ago

th-yong commented 4 years ago

Data augmentation issues related to adding parameters for rotation angles were identified in #4740

I thought of an augmentation, like (B), which I thought was the previous step in #4770

image

When the picture is rotated, the bbox is rotated by that angle, and I want to do an augmentation that recalculates and converts the values.

How hard do you think it would be to accomplish this?

WongKinYiu commented 4 years ago

All you need is segmentation label.

Babikoff commented 4 years ago

All you need is segmentation label.

Hello, I am sorry, what is "segmentation label", is it related to YoloMark, Darknet config or so?