Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How did YOLO resize images ? #464

Closed
phongnhhn92 opened this issue Mar 16, 2018 · 1 comment
Closed

How did YOLO resize images ? #464

phongnhhn92 opened this issue Mar 16, 2018 · 1 comment

Comments

@phongnhhn92
Copy link

Hello guys, I have a question, not a issue though.
I am wondering how did YOLO resize images ? For example, if the my input image's size is 1280 x 720 px and the input size of YOLO is 640x640 px. Then what kind a technique did YOLO use to resize not a square image to a square one ? In this case the aspect ratio between the original image and resized image will be different then how did YOLO resize the ground-truth bounding box as well. I hope you guys can help me to understand it clearly. Thanks

@AlexeyAB
Copy link
Owner

In general, there are 3 ways how to do resize - and each of them has advantages and disadvantages: #232 (comment)

For example, if all your training and test images have a resolution of 1920x1080, then they will be resized to 416x416 with identical distortions, so keeping aspect ratios is not required - detection will work fine.

I will add keeping aspect ratio.

sharowyeh pushed a commit to sharowyeh/darknet that referenced this issue Apr 9, 2018
zarmomin pushed a commit to zarmomin/darknet that referenced this issue Oct 1, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants