5 Best Free Image Annotation Tools

Andhika S Pratama
Data Folks Indonesia
6 min readJan 12, 2021


Image annotation is one of the techniques of labeling data for supervised machine learning. To do image annotation, one must need a dedicated annotation tool and there are a lot of image annotation tools right now.

Choosing the best one might be hard because many of the tools offer almost similar features but only different in some aspects. Dataset.com has listed all of the available tools that you can use and sort them based on who has the most star on their Github page.

In this story, I will give you the five best free image annotation tools based on my personal experience in using them for my projects or back then when I was searching for the most comfortable tools for me to use.

So, let’s get started…

5. imglab

imglab interface from imglab.in

imglab is the latest tool that I tried and it was told by my friend Ruben Stefanus who by chance saw the tool when surfing on the internet. This tool is a web-based tool but you can also install it locally. This is an advantage in itself because you could just access the website and start your annotation project. Also, no login whatsoever required.

imglab list of features

This is the list of features that imglab has to offer and it does what it says there. Pretty good tool with an easy to access website, very simple, and user-friendly interface.

4. VoTT (Visual Object Tagging Tool)

VoTT is an annotation tool from Microsoft that has several interesting features that differ from the other tools. The installation of this tool is very easy because you could just simply download the installer according to your OS from their Github page.

VoTT interface

VoTT offers active learning which is interesting but I never had a chance to use it. You could choose between Predict Tag and Auto Detect in the active learning feature. Besides image annotation, the tool could also be used for video annotation.

VoTT also supports many formats for exporting such as Azure Custom Vision Service, CSV, CNTK, Pascal VOC, Tensorflow Records, and VoTT Json.

There are many features that can be explored in VoTT for your convenience in annotating images. The downside for me personally is that the types of annotation in VoTT is limited to rectangle and polygon only.


CVAT interface

CVAT stands for Computer Vision Annotation Tool which developed by Intel. Besides image annotation, CVAT also supports video annotation too just like VoTT.

The people behind this tool also create a dedicated youtube video about CVAT but it is outdated now as CVAT has come with so many improvements as you can see by the interface alone.

What’s great about CVAT is the cleanliness of its interface while still putting in a lot of features. You can choose five types and it includes a cuboid which not many tools have and this feature called “AI tools” which again, haven’t got the chance to use it. As for the installation, well it is not the easiest tool to install because you need docker first.


labelimg was my very first tool in my image labeling experience. It was the very first time for me to get my hands on image labeling as my previous project or work was to annotate audio for speech recognition. The tool was recommended by my machine learning engineer and that he also gave me the guidelines on how to use it.

labelimg interface

For a first-timer back then, I was surprised by how easy it is to install the tool and how easy it is to start the program. Understanding the tool wasn’t that hard either due to the friendliness of its user interface. The drawback of this tool is that it only offers one shape which is a bounding box or rectangle shape. You can add another shape via coding on its GitHub page, but I’m no coder so I can’t do that.

labelimg also offers two types of files that you want to save your file into. The first one is PascalVOC and the second one is YOLO.

Still, it is a great tool for starters and if your project is relying on a bounding box only, this tool is for you.


This is the best tool that I currently use for my image annotation projects. labelme is more of the same as labelimg in terms of ease of installation and interface wise. The difference between them is that labelme has some of the features that make me use it as a daily tool for annotation.

labelme interface

One of the best features in my opinion is the “File List” on the bottom right. When you have a lot of images to annotate, there are chances that you might miss to annotate some of the images. This is why the “File List” comes in handy because it doesn't only list your files but it also gives a checkmark for each file that you already annotated.

In labelme, you will have the freedom to choose the six types that it has, starting from polygon, rectangle, circle, line, point, and line strip. labelme gives you the flexibility in annotating images while still giving you ease of use.

The only downside of labelme for me is that it could only save your file in JSON format. But that’s not a problem if the ml engineer is okay with the format.

That’s it, folks! the five best free tools for you to choose from!

Some tools may have better features than the others… but in the end, it comes back to your preferences and needs for the image annotations themselves.

Happy annotating, annotators!


Soo after trying out some of the best free tools here for my daily data annotation and giving it some sort of a review or opinion, I’m finally working in a data annotation service named Tictag.io.

I feel like I need to put it here because the tool is very interesting as you can do data annotation from your phone and get some rewards! you can basically do data annotation anywhere anytime!

You can check out the article that I made specifically for Tictag here: https://andhikasetiap.medium.com/data-annotation-is-now-at-your-fingertips-c195ede9489b

You can try it yourself and let me know your thoughts!



Andhika S Pratama
Data Folks Indonesia

Hi there! Currently, I’m a Data Annotator in Tictag.io who have an interest in writing such as Copywriting, and UX Writing.