<   SERVICES
Custom computer vision solutions powered by AI

Computer vision enables systems to identify objects on images or videos and take relevant actions. Using the most recent neural networks like GAN, we create reliable computer vision applications that allow our clients to automate work processes.

Get in touch
From detecting flaws in production lines to identifying people on social media, you can use CV in almost any industry:

Medicine

Analyze X-ray photographs and test results, automate examinations, Identify defective goods, control the use of personal protective equipment, monitor storages

Retail

Count customers and stock, monitor queues, optimize staff work schedule, make service prompt

Agriculture

Identify destructive insects, monitor crops, detect plant diseases, harvest faster and more effectively

Production

Compare properties, determine property value, manage documents easily, improve search results

Real Estate

Add AR and VR features to your app, recognize faces on pictures, create masks and filters

Insurance

Analyze and assess damage to property, reduce paperwork, automate risk evaluation

Finance

Simplify KYC verification, secure banking data with biometric authentication, prevent fraud

Our computer vision process

01

Idea discussion

We get in touch with you and find out the project details — what exactly you need and how we can meet the challenge. Then we offer several solutions based on your time and budget expectations.

02

Business analysis

It’s important to do business analysis at the start of any project: without it you won’t know if users need your product. We’ll study the market, competitors and potential users to help turn your idea into a successful product. After the research we’ll define the project scope and provide you with the quote.

03

Collecting data

Training a neural network requires a lot of representative images. During tech analysis we find out if we have enough data. If it’s not enough or we need images of better quality — we can find or create the required dataset.

04

UI/UX design

We create mind maps and wireframes to work through the basic logic of the app. Then we ‘dress up’ the screens with colors, fonts and animations — this approach helps us create beautiful and user-friendly interfaces.

05

Full-cycle development

We develop computer vision systems of any complexity for any business. Our development process is divided into sprints: this way we can show you intermediate results and keep you in the loop.

06

QA testing

Combining manual and automated testing, we make sure that your product works properly on all devices. We also run load tests to determine how much your computer vision AI system can handle.

07

Support after release

Of course, the work doesn’t end once the computer vision system is developed. We provide post-release support to all of our clients — all terms and details will be described in the SLA document.

What our clients say
Picture of the author
Roman Chistyakov
Founder of SAORSA startup

We worked with web and mobile app development company Brivian on the Saorsa project. The team was very good at communicating, responsive and friendly during the whole project. At all stages of development, the guys showed intermediate results and offered their solutions. Team took care to make the app convenient for the end user. They have extensive experience in mobile development and know many nuances which the startupper may not know about. This was very helpful.

Picture of the author
Derek Distenfield
COO, CFO & Co-founder of Gsdvs

I came with a cryptocurrency app request and Brivian was the perfect solution to build the utmost viable platform. They deliver excellent quality services and are customer-friendly. The team`s commitment to driving results and incredible technical expertise encourages long-term partnership. Highly recommend!

Our Insights
FAQ

Do you have enough expertise in computer vision applications?

We carefully choose tech specialists. Our experts have Master’s and PhD degrees in mathematics, specialize in different development areas and some of them have 20+ years of experience.

How exactly do computer vision, AI work?

Computer vision uses AI to train systems to see, interpret and understand visual data. Using digital images and videos and deep learning models, computers can identify and classify objects, and then react to what they “see”.

How many images do I need to provide for CV development?

Computer vision companies require a lot of images — about 1000 for each category. One of the first things that we do is evaluate the quality and quantity of the client’s images. If there aren’t enough images or they are of poor quality, we can help the client collect additional data. Find a required dataset, create a system that collects such data or request volunteers to make the needed pictures.

My project is very sensitive, how do I know that you won’t reveal my data?

If needed, we can sign an NDA document. This way you’ll be sure that we won’t steal your idea or disclose sensitive information.

How much time does it take for computer vision companies to develop a CV system?

We need about a month to develop an MVP model of such a system. The more features and requirements — the more time it takes.

How much does computer vision development cost?

To calculate the most accurate cost of our services, we multiply the hours needed for the project by the hourly rate of each specialist on the team. If you want to get a quote — feel free to contact us. Just fill in the form below 😉

What is your pricing model?

Brivian works based on the Time and Material model. It means that you pay for the working hours of our specialists and materials used for the project.

Who will manage my project?

Every project team includes a project manager who will be responsible for communication and make sure that the team meets the deadlines.

Which neural networks do you use?

We apply the most recent technologies, such as generative adversarial networks (GAN). They appeared not a long time ago, and most computer vision companies still use old technologies and don’t maximize their results. We do because GAN networks show a higher output.

Which technologies do you use?

Usually, we create computer vision systems using Python. When developing systems that need fast object recognition, we code in C/C++ — for better optimization. Here’s what we use most often:

Languages: Python, C/C++, Go
Libraries: TensorFlow, Keras, OpenCV, PyTorch, SciPy, Pandas, NumPy, Spark, Chainer