Classifai v1.1 is released!
Classifai, a browser-based data annotation tool to prepare labelled data for supervised learning, just had a release. This brings new features and bug fixes which would ease the image annotation process even more.
How to Install Classifai?
Classifai comes with installation packages for operating system as shown below.
Choose the installation package from here accordingly to install Classifai. Its just a few installation steps away to use Classifai.
Alternatively, Classifai is built with Java, which makes it platform-agnostic. With Java 14 installed, download the jar file and run it with the command below
java -jar classifai-uberjar-1.1.0.jar --unlockdb=true --port=9999How to run Classifai?
Classifai comes with desktop icon. Double click it will prompt the starting of the tool.
Classifai has in-memory database. This allows annotations to be stored in the system. This is a major value add as often the dataset to be labelled is humongous and a project for data annotation likely takes a few days.
What are some of the new features?
Project Deletion Feature
Back in the initial release (v1.0.0) of the tool, there’s no deletion feature. We added that in. Now deletion of existing projects is just two clicks away.
Image Data Deletion Feature
To remove images data imported to an image project, simply click on the images panel and proceed with selecting images to delete.
Note: The selected images will only be removed from the project while stay untouched in the stored path of the system.
Segmentation Technique — Drawing while Shifting Image
One motivation for segmentation approach is to define the exact silhouette of the object of interest. The ability to draw while shifting the image makes the process more doable, focusing on every area of the object outline.
This post illustrated through some of the significant features of the new release. For a complete list of the changes including bug fixes, check out the release note.
What’s next?
Up next, Classifai team is preparing blog posts to talk about how to prepare quality labelled data, as well as walkthrough of AI modelling process with the customised dataset.
Contribute to the development of Classifai on Github. Most importantly, we welcome you to file any issues or feature requests of the tool for us to best support you.
Check out the resources below for more information:
