Yolo_Autolabel

Requirements 📋

Python 3.6 or later.

To install run:

pip install -r requirements.txt

or simply

pip install opencv_python
pip install PySimpleGUI

How to use

Run:

$ python autolabeler.py

Now select:

This will produce this output:

The images with the detected objects (if 'Save image' is selected) and the labels for every image (if Save label is selected)

Now to export the dataset:

This will produce this output:

export
|
+--obj_train_data
|  |
|  +--soccer.txt
|  |
|  +--dining_table.txt
|
|  +--... (rest of the labels)
|
+--obj.data
| 
+--obj.names
| 
+--train.txt
| 
+--images.zip
| 
+--upload.zip
| 
+--labels_to_cvat.txt

images and upload.zip files are ready to be uploaded to CVAT if you need to edit the labels or export into another format.

How to upload labels to CVAT

Create project

copy labels_to_cvat.txt into raw and click Done

submit

create task and upload images.zip

submit

Open the task and click Upload annotations YOLO 1.1 and upload the upload.zip file

ready to go!

Download yolov4.weights file 245 MB: yolov4.weights

Autores ✒️

• Oscar Mauriaca - Desarrollo - xaerincl