What is Wizart Vision API and how does it work

Product Updates
Kseniya Tarasevich
5 mins

Wizart presents its Vision API. What is it? How does the technology work? And how can the user install it? Find the answers below. 

 What is Wizart Vision API? 

“We are proud to say that no other AR-visualization tools on the home decor market can compete with Wizart tools. We are the only solution able to do multi-surface visualization on ceilings, floors and walls at the same room simultaneously. Moreover, the Wizart visualiser does the best regarding vertical surfaces. Our RnD team has gathered enough expertise and keeps developing new directions in the computer vision field”, — Pavel Batashov, the Wizart CTO explains.

Indeed, Wizart engineers work on various computer vision solutions for the interior design market. They managed to create excellent visualization, segmentation, and detection of the shadows and lighting points. The neural network can detect the geometry of the room and wall quantity. 

"We are at this point when our technology can accomplish much more than we use, and we can easily fill up the other needs of the market", - Pavel Batashov says. 

Why Wizart Computer Vision API is valuable

Wizart Vision API has many unique features, that made this product unique at the market. Firstly, it provides a wide range of applicability. 

“Clients use ready-made solutions, but they are not limited in possibilities to implement their own ideas. Feel free to create a completely custom product that will meet exactly your business needs and show how unique your brand is”, — Pavel Batashov says. 

Secondly, our team guarantees secure and stable work. Dozens of monitors and an incident management system, as well as secure hosting from Amazon. All of those components are designed to avoid many troubles from our customers. 

“And we are proud to provide the flexibility. It starts from our pricing policy and goes to the ability to make changes to the roadmap according to your requests”, — Wizart CTO comments. 

​​Creating one working model requires a lot of effort, while Wizart has dozens of them. You will need to:

  • describe the business task in detail, in order to develop an understanding of the conditions in which the model will work
  • outline which data will be required
  • how the model will be validated for performance

The next step is the relevant data set for learning and testing. When we say “relevant data set” we suppose that this type of data will be uploaded to our model. Nevertheless, it is extremely hard to get enough data sets of the required quality. 

“We expect dozens of thousands of those data to create renders of such impressive quality. We also managed to work up the uploading rules, but we continue to work on the marking regulations and testing our data quality. And I have to add, this is an incredibly difficult and long process, which is required even to get the one segmentation model. This is why it is easier to use ready-made solutions”, — Pavel Batashov explains. 


Indoor semantic segmentation provided by Wizart Vision API, the graphic is created by Wizart team

How to run the Vision API

Wizart shares access to the solution via REST API, accompanied by comprehensive documentation. 

It works in three simple steps: 

  • fill out the sign-up form 
  • get the API token 
  • use the solution

Interested? Find the details here.