how to download labelme for image segmentation tasks
Today I will teach you how to download labelme, the most widely used tool developed by MIT to annote images for object detection, computer vision researches. We normally will use labelme to segment our images to obtain a certain dataset or json file. Below i will teach you how to download and use labelme step by steps:
- the environment we will use is anaconda3. First we make a new environment. We go to the terminal and type in the following command:
conda create -n labelme python=3.6
2. then we will need to activate the labelme. We type in the following command:
conda activate labelme
3. now we can start downloading, type the following code in the terminal:
conda install -c conda-forge pyside2
conda install pyqt
pip install pyqt5
pip install labelme
4. after downloading all of the above packages, type in “labelme” in the terminal and the software will run and bring you to this page.
5. the software is very user-friendly and we can start uploading image by clicking on OpenDir on the left-top. Here as an example, I will open an image of MJ because the “last dance” is on.
6. Now we can annotate the image by clicking on the “create polygons” on the toolbar on the left. Don’t forget to give the polygon a name and save it :)
After saving it, just click “next image” to continue marking other images.
Hope this tutorial is helpful and comment below if you have any questions
For your reference, the link to the labelme github is
