If you're looking to develop accurate and effective machine learning models, data annotation is an essential part of the process. Fortunately, CVAT provides top-quality data annotation services that are ready to annotate your data. In this article, we'll walk you through the step-by-step process of ordering data annotation services from CVAT.

‍

Let’s start with requirements:

‍

• The data must be visual: 2D images/video clips or 3D point clouds.
• The minimum budget for a project starts from $10,000.

‍

If you fit them, then the process will include seven steps:

‍

Step 1: Submit a Request

‍

To start, you should send us your requirements by email: contact+labeling@cvat.ai. Tell us the type of annotation required, the volume of data, and the turnaround time.

‍

Step 2: Trial annotation

‍

After we receive the request and process it, we will ask for a trial data sample and initial specification from you, to better understand your needs and the way the dataset should be annotated.

‍

We may sign an NDA if needed.

‍

After we receive a trial data sample, we will start annotating it. From the beginning to the end of the process we will communicate all our questions and concerns that may arrive. For example, how data must be annotated, if you have an example of desired annotation, how to annotate small objects or when one object is behind the other. 

‍

We will update the data annotation specification based on the answers.

‍

After the data annotation specification is ready, we will show it to you. Specification must be approved by both sides to continue with the initial proposal.

‍

Step 3: Initial proposal

‍

If all is fine with specification, we calculate the time and cost for the annotation of the full set. In some cases the final price for the project will be known if we can calculate the number of objects in advance. Otherwise the price will be per one data annotation object (e.g., bounding box, skeleton, point, polygon, 3D box, etc).

‍

Step 4: Signing the contract

We will sign the agreement and  include the specifications as an appendix. Once this is complete, we will await the transfer of data to begin the annotation process.

‍

Step 5: Annotation

‍

Once you transfer the data to us, the team of experienced and qualified annotators will annotate the data according to the specifications we agreed upon. 

‍

Step 6: Quality Assurance

‍

After completing the annotation process, the CVAT team conducts a quality assurance check to ensure the accuracy and consistency of the annotations.

‍

We will use the statistical methods that have been selected or agreed upon in the contract.

‍

Step 7: Delivery

‍

Finally, the annotated data will be delivered to you. Together with annotations quality report and description of used quality assurance methods.

‍

If any inaccuracies are discovered during this stage, we will provide free fixes according to the estimations that were agreed upon in the contract.

‍

In conclusion, by following these simple and effective steps, you  can obtain high-quality and accurate annotations for your machine learning models through CVAT and improve the decision-making processes and gain a competitive advantage in your respective industry.

‍

Any questions left? Feel free to contact us!

Discord

LinkedIn

Gitter

GitHub

‍

‍