Simple GUI that lets you annotate points in images and capture their coordinates.

A .txt file is exported containing each point's label and coordinates. Specifically, it has the following format:

# Points2D.txt
img_name point_label x y
<name_of_img_1> <label_of_point1> <x coord> <y coord>
<name_of_img_1> <label_of_point2> <x coord> <y coord>
...
<name_of_img_N> <label_of_pointM> <x coord> <y coord>

Additionally, the annotated images are saved in the output directory.

Installation

You need a python version that supports tkinter (python 3.6+). In Ubuntu, you can install it with:

sudo apt install python3-tk

Then, install the requirements:

pipenv shell # To create the virtual environment
pipenv sync # To install the libraries as specified by the Pipfile.lock

Usage

Within the newly created python environment, run main.py with the following arguments:

• path: The path to the folder containing the images to annotate.
• (optional) --output: The path to the output directory where the points2D.txt and annotated images will be saved. Default: ./output

For example:

cd point-coords
python3 src/main.py ./test_imgs

After running the command, a window will appear with the first image in the folder. COntrols are:

• Left click and drag: Move the image around
• Scroll in/out: Zoom in/out
• Right click: Add a point
• Next button: Go to the next image

After right-clicking, a window will appear asking for the label of the point. After entering the label, the point will be added to the image.