Computer vision with an eye For Real Estate

Project Information

CV4RE

Grant agreement ID: 823176

DOI

10.3030/823176

Project closed

EC signature date 3 August 2018

Start date 1 September 2018

End date 31 July 2020

Funded under

• INDUSTRIAL LEADERSHIP - Innovation In SMEs

Total cost

€ 1 695 653,75

EU contribution

€ 1 186 957,63

Coordinated by

Mercurio Platform SL
Spain

Periodic Reporting for period 2 - CV4RE (Computer vision with an eye For Real Estate)

Reporting period: 2019-07-01 to 2020-07-31

Summary of the context and overall objectives of the project

CV4RE is a project from Restb.ai a company that started its activity in 2015 with a proprietary technology based on Deep Learning Convolutional Neural Networks that allows machines to understand, categorize, classify and tag images, with accuracy rates reaching 99%.

When looking for a new home, it is very common nowadays to learn about neighbourhoods, nearby schools, and facilities, and even the potential future value of a property, before approaching a broker. However, once the initial search criteria are established, the process has not changed a great deal. The traditional way would involve agents spending hours searching a database, flipping through properties and pictures to find a suitable home that would potentially satisfy their client’s home buying needs – then weeks or even months exchanging links and information with clients, and maybe even showing a few properties at the end of this lengthy process. Restb.ai is looking to address this very issue by helping online real estate portals effectively fast-track this process by providing a revolutionary online browsing experience, using Computer Vision.

Main objective

Thanks to the CV4RE project, Restb.ai aims to achieve €16 M turnover by 2024, with NPV€10 Million and payback of 2.23 years (considering the 2 year of project/development where there are no profits). The willingness-to-pay for CV4RE is assured by the interest we have received from our potential customers and all the actors of the value chain that will benefit from our solutions. We have already signed a contract with Idealista (online RE portal leader in Spain), Apimo (Major CRM Company in France) and we are in contact with IKEA that could give us the maximum visibility in the Home Furnishing sector.

Conclusions of the action (final period)

We finish the CV4RE project with all our goals achieved and very proud of the hard work done by all the team.

The technical development, as usual in these cases, has been harder and a bit longer than expected, but thanks to the hard work that everyone put together and the talent of our team, we have been able to achieve all the main technical goals.

Technical development:

Being a start-up that focuses on state-of-the-art technology based on Artificial Intelligence it is really challenging. Because of that we are always trying to iterate and push the boundaries of what its possible in terms of technical developments ands products for our clients without forgetting about our Platform, our internal tools, and our pilot programs.
After concluding CV4RE, our Platform has now new capabilities that enables clients to consume and use our products much more efficiently and easily, having the opportunity of process their images from and through every part of the world.

Our technology has been enhanced with new powerful algorithms and new solutions that will help us make a difference from our competitors in the markets we compete.

Our Internal and Sales performance tools have helped us dramatically, increasing the efficiency of our internal processes and tasks, and reducing a big portion of the time we were spending on them. Our ability to show to the clients our solutions and conduct POC has been increased meanwhile the integration of our API its now much easier than it was before.

And finally, been able to pilot some of the new iterations of beta releases with some of our clients gave us much more information and knowledge of what the market needed from us, which improved our products greatly.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The work performed around this project has been:

WP1: API Platform new capabilities

Our Platform has been strongly enhanced thanks to the new capabilities developed. In particular, the Multi Region Platform Deployment development has helped us to have a much more robust and solid platform, with a greater degree of redundancy and availability.

WP2: Enhanced AI

This has been the most important work package of the project. In this work package we have improved, expanded, and enhanced our core AI technology meanwhile building new algorithms that helped us providing new solutions and products to our clients. Almost all the company has been involved in the development of this work package. Starting from the marketing and sales team, getting as much information from our clients as possible on how the final product should look like and which functionalities would make more sense, followed by the Product team converting that information to the Engineering team in terms of requirements and specifications that finally the Tech and Data team materialized in the development of the new solutions and the enhancement of some of the old ones.

WP3: Internal Tools

Thanks to these developments our internal tools have been greatly improved in terms of usability and efficiency, not only for our teams, but for our clients too.

WP4: Sales Performance Tools

As a result of this, clients understand better what we do when looking at out website, interact and test oud Demo in a much simple way and integrate our solutions using less resources and in less time that before.

WP5: Piloting:

As a result of this, clients have been able to test the early versions of our new solutions, giving us important and critical feedback of what was missing or the things that could be improved.

WP6: Commercialization and communication activities

The Business side of the company has been greatly improved thanks to all the actions that we have taken, like improving our presence and branding in the Real Estate sector, finding new sub-markets or niches inside Real Estate interested in our solutions, having presence in more and bigger conferences and events and promoting our products there by doing presentations or having stands where clients could approach us, and last but not least by having a stronger presence in the media (websites, 3rd party blogs, campaigns, social media, audio visuals, press and more…).

WP7: Management

As a result of the management activities done, we have ensured the exit on the achievement of the technical and business goals, by managing the resources and teams of the company to achieve Milestones and to complete tasks before the deliverable dates of those.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Computer vision identifies, categorizes, and correlates features in and across photographs allowing businesses to embrace the image-heavy digital era and unlock a treasure trove of business-critical information.
If a picture is worth a thousand words, then computer Vision unlocks access to a wealth of impressions, moods and emotions that make potential buyers and renters act on a property, including, and especially, the blind.

We are the only company that has built its proprietary computer vision technology from the ground up to specialise in real estate, to understand and describe property photos with metadata that are specific to real estate. The cloud-based technology is highly scalable and can easily grow to process billions of images.

Thanks to the CV4RE project, we have been able to improve our solutions, improving our algorithms, the architecture of our models and increasing the amount of training data.

Providing our customers with such precise and relevant metadata returned from our solutions, gives them important benefits, that materializes in the increase of traffic to their sites and higher conversion rates.

Restb