Select one of the four FAQ sections:
What is Zillin?
Zillin is a web tool for annotating images and managing image datasets. Most often it is used to create training data for big neural networks in computer vision.
What can I do with Zillin?
You can upload image datasets, configure marking tools, attributes and questions, and do the actual annotation work. You can do this yourself or invite other people to cooperate. Then, after all the work is done you can export your annotations and use it for training or testing your computer vision algorithms, most typically ones that are based on deep neural networks.
What is a dataset?
A dataset in Zillin is a single identifiable collection of images. It may be one batch of images acquired on a particular date, or a collection of cases showing a particular type of defect. Or whatever you consider relevant. What is important, when you publish a dataset it becomes frozen. Modifications are highly restricted then to make sure your experiments are always repeatable. Multiple datasets can be added to one project and one dataset can be added to multiple projects. A dataset itself does not include annotations.
What is a project?
A project in Zillin consists of three things: (1) a selection of image datasets, (2) a configuration of marking tools, attributes and questions, and (3) the actual annotations. You create a project to make use of our datasets and prepare training data for your deep learning models.
Do I need to be a scientist to use Zillin?
Zillin is for everyone who needs to work with image datasets. The product has been created to simplify image annotation. Creating and configuring a project is pretty straightforward.
Can I use Zillin to simplify annotation of large datasets?
You can invite human Annotators to perform image annotations. When you invite someone as an Annotator, he or she will receive an email invitation with a link to the marking view of your project. After successful login to Zillin, the Annotator will be able to annotate your dataset image by image, using marking tools or answering questions you have made during configuration. At the end of annotation in a particular image the Annotator will have to submit his or her work. Otherwise, if the Annotator will be unable to annotate image, he or she can simply skip it and move to another image.
Then, there is an optional review process. After an Annotator has finished work you may conduct review to double check if annotation has been done correctly. A Reviewer who has been invited by you also receives an email invitation. Images which passed the marking process are waiting for review. The Reviewer can start reviewing them image by image. The Reviewer will see all work made by an Annotator, all selection of elements, all answer for questions (but without information which Annotator has made the work). The Reviewer will have three options to do:
After the work is finished, you can export all data and use it to train your models.
How to start using Zillin?
First of all, you need to register an account in the Zillin platform. After successful logging in, you will be able to add datasets and create your first project and check out how it works. Just keep in mind that you first need to add and publish datasets before creating your projects.
Where is my data stored?
Your data is securely stored in the Microsoft Azure cloud so that you don't need to worry about it and you can access it everywhere. Private instances of Zillin are also available as a custom service. For maximum security you can also enable end-to-end encryption, so that your images are 100% unreadable in the cloud (available in the Enterprise plan).
Is my data secure and encrypted?
Yes. Zillin uses a variety of proven security controls to ensure your data is secure including login authentication procedures, transmission encryption, firewall barriers and - most notably - end to end encryption (available in the Enterprise plan).
Who else can see my data?
As a workspace Owner you decide who gets access to your data and you control the permissions with Zillin's predefined user roles. Our employees do not have access to your datasets and projects unless you explicitly add them to your Workspaces. On the other hand, due to strictly technical reasons a limited number of our administrators may theoretically get access to raw data on the servers in the cloud. If you want to remove such a possibility turn on end-to-end encryption of images (available in the Enterprise plan).
Can I store data protected with an NDA (non-disclosure agreement)?
First of all, invite your customer to use Zillin together with you. Very often they will be happy to exchange images through a secured and trackable channel. If needed, turn on end-to-end encryption so that the images are fully protected – even against someone with physical access to the data center infrastructure. Your NDA will be fully fulfilled as no image data will be accessible without your encryption key.
How does end-to-end encryption work in Zillin?
We use a symmetric AES-256 algorithm for encrypting your image data BEFORE they are sent to the cloud. When you create a new workspace, you have an option to enable encryption. It will cause creation of an encryption key. This key will be created in your browser and it will never be transferred to the cloud. You just save this key and you can share it with your team members using any secure channel of your choice. Every time you enter that workspace, you will be asked for the key. Again, this key will only stay in your browser's memory and it will be used to decrypt image data coming from the server.
NOTE 1: We encrypt all image data, but not meta-data – the project names, project configurations, e-mail addresses of team members and annotations cannot be encrypted, because these are processed on the server side (while images are processed only on the browser side). If you consider this information also extremely sensitive, we suggest that you use artificial project codes and synthetic e-mail aliases (e.g. firstname.lastname@example.org).
NOTE 2: End-to-end encryption is available in the Enterprise plan.
Can I work with images or annotations that contain personal data?
Due to the European law and as explicitly stated in § 10 of our Terms of Service, you will be required to obtain our consent and enter into a relevant agreement with us about processing personal data. However, after enabling end-to-end encryption you can upload images containing personal information without any concerns.
Can I have a private instance of Zillin?
Yes, this is possible through a custom service (subject to additional charges). You will need to create your own Microsoft Azure account and we will provide you with a docker container with the Zillin application. Further upgrades will consist in replacing the container with a new one. Please contact our sales to discuss details.
Can I use Zillin with another data storage provider?
Switching to a data storage provider other than Microsoft Azure will require customization of Zillin. It may be possible as a custom service.
Can I use Zillin on my own local servers, without anything being stored in the cloud?
Yes, although some further customization of Zillin will be required. Please contact our sales to discuss details.
How can I set up a project in Zillin?
Before you start creating a project, make sure you have some datasets already uploaded and published. Then, click PROJECTS → CREATE A NEW PROJECT, and:
How can I use marked data in Zillin?
All work that has been done in your project can be simply exported as a JSON file. Just go to your project details → Export tab, and you will be able to download the file from there. This file will contain all your annotations together with additional meta-data such as who and when created each marking.
Who can work on my project?
You can invite anyone by providing an email address and selecting a role for this user.
What roles are available in Zillin?
In Zillin we distinguish several access-control roles with gradually increasing permissions:
Can I add a new team member during the work on a project?
Yes, you can. Go to the project details and invite a new team member.
How can I delete a project?
You can delete a project from PROJECTS view. There is an option for that in the dropdown menu opened by clicking on the three dots icon next to the selected project name.
How can I delete my datasets?
You can delete a dataset from the dataset view by pressing at a trash icon or by going to dataset details and pressing the DELETE button. If the icon or the button is inactive, it means that you assigned the dataset to some project. To prevent losing your work we deactivate this option here. However, you can always go to project details, unlink the dataset from the project and then come back to datasets - it should be possible to delete the dataset now.
Why can't I add more images to a published dataset?
We decided that only a Manager should be able to do this. Developers and Collaborators may add images to a dataset as long as it is a Draft. However, after the dataset becomes Published there come some restrictions. Our aim here is to avoid accidental disruption of work in projects that are already using these datasets. Anyway, if you collect a new batch of images we recommend that you create a new dataset for them.
Does Zillin provide support?
Yes, you can contact us by email . We will try to help you as quickly as possible.
What data format can I export?
You can export your data in a JSON format. If you are interested in other formats, please let us know.
Our format documentation can be found here.
Does Zillin have an API?
Yes, it has, read it now - API Documentation.
If you can't find an answer, contact us and we will be happy to help.
+48 32 461 23 30
We are happy to help, 8:00-16:00 CET.
The best way to get answers fast.