Since we launched Roboflow in early 2020, our vision has always been to improve and streamline the workflow of computer vision projects so that developers can focus on the parts of their project that are unique, not on reinventing the wheel. But there's been one part of the process that has remained particularly cumbersome: labeling images.
Since the beginning, Roboflow has supported importing labels and annotations from all of the most popular labeling tools (including CVAT, LabelImg, VoTT, VGG Image Annotator, and LabelMe). This is great for teams that are extending an already established flow, but, if you're starting a new project from scratch, having your labeling tool separate from the rest of your pipeline adds a ton of friction.
Label images within Roboflow
Today, we're launching Roboflow Annotate, a self-serve image annotation tool built right into Roboflow.
We designed Roboflow Annotate to meet the following needs:
- Integrated directly into the workflow so whether you're labeling an entire dataset from scratch or correcting a single errant annotation it's simple to jump in and out of the flow.
- Easy to collaborate with your team so you have a single source of truth, not a collection of divergent datasets on various team-members' laptops.
- Simple, streamlined labeling interface with smart defaults to help you label quickly and accurately.
Built with extensibility in mind
By integrating the labeling flow into the rest of the computer vision pipeline it enables us to do some unique things that are hard to do if your tools aren't designed to work together.
For example, we're already piloting model-assisted labeling with a few select partners so you can use the state of the art models you've trained with Roboflow Train to jumpstart your labeling. And we've seen customers getting monster benefits from active learning by integrating our upload API into their applications to feed production data back into the system to be labeled and fed back into the training pipeline.
Contact us if you're interested in being an early beta tester of these advanced labeling integrations. And stay tuned because we have a few tricks up our sleeves to further improve Roboflow Annotate in the months ahead.
Try it out
Roboflow Annotate is included with all of our existing plans for no additional charge. You can try it with your datasets today; simply click a source image in the interface to enter labeling mode.
And be sure to check out the Roboflow Annotate docs to learn about the different tools and keyboard shortcuts available.
Cite this Post
Use the following entry to cite this post in your research:
Brad Dwyer. (Dec 8, 2020). Introducing Roboflow Annotate. Roboflow Blog: https://blog.roboflow.com/labeling/
Discuss this Post
If you have any questions about this blog post, start a discussion on the Roboflow Forum.