drawyolo 0.1.1

Draws boxes on images from annotations in YOLO format

Draws boxes on images from annotations in YOLO format.

Installation

Install using pip:

pip install drawyolo

Or directly from the repository:

pip install git+https://github.com/rbturnbull/drawyolo.git

Drawing boxes from labels

If you have a label file in YOLO format like this:

class_id x_center y_center width height

For example:

0 0.35 0.35 0.14 0.14
0 0.55 0.3 0.14 0.14
1 0.7 0.15 0.2 0.2

Then you can draw the boxes on an image like this:

drawyolo image.jpg output.jpg --labels labels.txt --classes class1,class2

To try it out on the image in the repository:

drawyolo tests/test-data/terrier.webp tests/test-data/output.jpg --labels tests/test-data/labels.txt --classes eye,ear

That will create an image like this:

To resize the image, use the --width and/or --height options. The aspect ratio will be preserved if you do not set both --width and --height.

For example:

drawyolo tests/test-data/terrier.webp tests/test-data/output-thumbnail.jpg --labels tests/test-data/labels.txt --classes eye,ear --width 240

The thickness of the line will be set according to the final size of the image. You can change the thickness with the --line-thickness option.

Drawing boxes from a YOLO model

If you have a YOLO model with weights, you can draw the boxes on an image like this:

drawyolo image.jpg output.jpg --weights model.pt

You can also resize the image as before with the --width and --height options.

Advanced usage

For more options see the help

drawyolo --help

Credits

Robert Turnbull (Melbourne Data Analytics Platform) - https://robturnbull.com