Private and secure data
Useful Machine Learning (ML) model needs big, high-quality data. Zillin team put years of experience in data preparation creating a tool that makes the process fast, simple, and faultless.
You can do your project on our own or build a team with your colleagues. Simply type in their e-mail address and select their role in your project. Read more
Right now two roles are available:
Classifier - if you want someone to do annotations for you, invite them as Classifiers, which can complete the work and submit it to you.
Reviewer - if you want team members to do the quality assurance, invite them as Reviewers, which can either reject the annotations, correct it or confirm.
Zillin offers right now three tools to annotate images. Read more
Bounding boxes - which is probably the fastest way to segment the right content on your images. With this tool, you can select most of the objects.
Polygons - it is the most precise way to select objects on your images. With this tool, you can precisely select the contour of the object that you want to segment.
Questions - allow you to ask for information about elements, where simple selection is not enough and you need to go more into detail about specific aspects of the dataset.
This tools can be used each separate or together, everything depends on the result you try to achieve. More tools are coming soon.
During the process of image annotation, mistakes can and probably will happen, which could affect the quality of your dataset. Read more
That's why we have added the review mode, with is an additional quality assurance layer as a possible last step of the process. Reviewers can check the quality and reject, correct or accept the work that is done. The review is an optional step and is not required, you can always export the dataset as it has been annotated.
Working on a dataset with thousands of images requires sometimes external workforces. Read more
Zillin will create a team of image annotators for your project that will fit your needs. Our team of experts will analyze your project requirements and set up the complete workflow making sure you get your desired results as fast as possible at a reasonable cost.
Each selection your team makes on the satellite images or aerial photography (also created by drones) will automatically include the exact coordinates. This feature is handy when preparing data for sea vessel tracking, infrastructure damage analysis, soil classification and crop detection, or many other popular ML projects.
We know that some things might seem complicated at the beginning. That’s why we listed the most common questions in the section below. Here you should find the answers that you are searching for. If your questions are more detailed, feel free to contact us at any time.
On the project view press "Create New Project" button, then in 3 simple steps configure all project:
that's it, your project is ready for publishing or you can simply preview your project before publishing to see how your people will see your project.
All work that will be done in your project can be simply exported in JSON file. You need to go to your project details → the export tab, and you can download the export file from there.
For your comfort we stored all data in our storage, so you don't have to worry about it. Although, if you have your own storage you can contact with us, we will try to sync.