LabelImg - HumanSignal/labelImg GitHub Wiki

Filename:labelImg.py

Pre-requisites

• pyqt5==5.10.1
• lxml==4.6.3
• Anaconda

Description

• This is a tool used for data annotation. Create new lables or show the existing lablels into the image (annotation visualization).

Note

• If you don't know how to create an conda environment refer-To create conda env
• To Understand how to use the tool refer-Labelimg
• For annotation visualization: copy the existing labels to same folder with the images

Output

• Creates a txt file in the yolo labeling format which contains annotation corresponding to the images which has <object_class_number><x_center><y_center><width><height>

Steps to execute

• Create a conda environment
• Run the command python3 labelImg.py

Image of the output

---

Steps to correct the labels

• Change the “predefined_classes.txt” to the classes you want in the data folder
• Click on “open Dir” and open the folder where you have all the images
• Put the classes.txt which has all the classes inside the label folder
• Click on “Change Save Dir” and open the folder where you have the labels
• Correct the labels and save