What are the product specifications?

All product specifications are listed on the product page you will access after selecting one of the pre-made datasets from the landing page of this site. If you have questions regarding a pre-made dataset please contact support.

What is the difference between an organic and synthetic human dataset?

Organic datasets are photos taken of real humans. All individuals have signed a biometric release that allows the use of their image and biometric data for machine learning purposes.

A synthetic human dataset consists of images generated by AI. Since these people do not exist, the use of them does not fall under privacy regulations.

How many people are in the organic human dataset?

We are constantly adding new images, however, as of July 2022 there are over 1500 people with over 500,000 images from different angles, and with different expressions.

The demographic distribution of the organic dataset can be downloaded from the dataset product page.

What can the human datasets be used for?

The human dataset can be used for all machine learning purposes. All rights for the trained model are transferred automatically to the purchaser.

Using the datasets for training generative models for visual synthetic media require an extended license. This license adds an additional cost per image that is detailed on the Pricing page.

How are the datasets tagged?

We have basic standard text to describe the image to a medium level of detail. These tags are delivered alongside the dataset as text files with a corresponding name to the original image file.

Biometric datasets contain standard tags including facial expression, emotional expression, head pose, hair style, age group, gender and ethnicity.

During checkout you can modify the standard tags of your dataset. You just select the Tag option before making your purchase. Custom tagging is charged at a flat rate. You can view the rates on the Pricing page.

What can I do with the purchased datasets under the standard license?

You can use the datasets for any machine learning purposes with the exception of synthetic media. All rights, usage rights and copyright of the trained models are according to the EULA automatically transferred to you. This includes the extraction of biometric data - for example, for facial recognition or emotion detection models.

You are not allowed to pass any of the dataset images themselves or any contained files to any third party. Under no circumstance can the dataset be published, copied to third parties or in any way shared.

What can I do with the purchased datasets under the extended license?

Anything you can do with the standard license plus training generative models which create visual synthetic media.

You have full usage rights to use the output of your trained model for any commercial or non-commercial purpose as long as it doesn't violate any terms of our EULA.

Can I use the purchased datasets be used for commercial purposes?

Yes. By purchasing one of our datasets you get the commercial usage rights to use them in your machine learning projects.

Why is my country missing from the dropdown in the account setup?

In order for us to comply with international laws we have removed some countries from the list available to add billing information. In support of the Ukraine we observe trade embargoes on Russia. Regretfully, the sweeping laws allowing data scraping in the UK presents a risk in selling datasets to UK entities.

Can I upload the dataset alongside my public open source project on sites like github?

No. Any publishing of the dataset or any transfer to a third party is not allowed.

How much does the one time purchase of a datasets cost?

The price of the datasets vary according to factors such as image resolution, quantity of images and how you intend to use the datasets. When you choose your dataset you will be presented with a range of pricing options that will determine the final price in the shopping cart.

Augmentation and custom tagging incurs additional charges.

View our pricing page for further information.

How much does a subscription cost?

Optionally you can purchase a subscription for your product. If you purchase your initial dataset with a subscription, you get the first year of updates for 30% of the purchase price. If you decide to subscribe to updates to your dataset after the initial purchase, the regular price of the annual subscription is 50% of the purchase price.

Can I subscribe to custom build datasets?

You can currently only subscribe to our pre-made dataset products.

What is the difference between a one-time purchase and a subscription?

We constantly add material to some of the pre-made datasets and upgrade them on a monthly basis.

If you purchase the one-time license, you will get all the material of the corresponding produced up to this moment.

If you purchase the subscription, you will get monthly updates where we add new material to your dataset.

What if there is no subscription option in my product?

Some datasets will not be regularly updated, and in those cases, no subscription option will be provided.

Can I download my dataset any time?

You can download up to three times per month. Contact us if you need to request more.

What payment options are available?

You can use Stripe to pay using credit card or ACH (bank transfer).

How do we delete our account?

You can edit or delete account in the Profile section of the site. If you have any issues please contact support.

Can we have more members logged in from one company?

You can have multiple team members login using the same account. Charges are by dataset, not per seat.

How do I see past invoices?

In My Account (when you are logged in) you can view past invoices in your Products section.

How do I change the password?

In My Account section (when you are logged in) you have the ability to change your password under My Profile.

Where are the images in the datasets obtained?

The datasets consist of images that are either fully-owned by vAIsual Inc, or have been acquired under a legal license from third party providers.

Is your synthetic facial imagery data sold as non-PII?

Yes, we sell synthetic datasets that contain AI generated images of humans. These datasets do not fall under PII (or GDPR) compliance laws.

Consequently, the legal consequences for the purchaser are drastically reduced.

What releases are used for the organic human dataset?

All of our real-life models are required to sign GDPR compliant model releases. Each model has an opportunity to read the model release and have it explained to them by our staff prior to the photo shoot.

Can we view the consent form signed by each model in the vAIsual human dataset?

Yes. Please find the relevant model release on the dataset product page.

In the consent form (signed by each model in the human dataset), does it share the exact purpose under which the data is collected?

Yes, our model release explains in detail what the photography will be used for. Prior to each photo shoot, every real-life model has this explained to them in detail - including the full transfer of rights to the images and biometric information.