Inoveo3D unveils its first 3D reconstruction results from simple photographs

Reconstructing a real environment in three dimensions today relies on equipment or processes that quickly become constraining: 3D scanners, LiDAR sensors, specialized acquisition, or photogrammetry requiring large numbers of images and substantial processing time. At Inoveo3D, we chose to explore a different path. Our ambition is to develop a technology capable of reconstructing a 3D environment from simple photographs, while reducing dependence on equipment as much as possible. The goal of our work is therefore not only to improve existing 3D reconstruction, but to profoundly simplify how it is used.

A 3D reconstruction is not an end in itself. It becomes truly interesting when it can support new uses, and that is where our other technologies come into play. In real estate, we can imagine connecting a reconstructed environment to online furniture catalogues in the future. A user could rebuild their apartment from their phone, remove or modify certain elements, then see furniture from a partner retailer's catalogue placed directly in their environment. The passage between the real world, a 3D environment and online commerce would become far more natural. In property development, this technology could also make it easier to create immersive experiences or personalization tools. But the potential goes well beyond this sector: architecture, industry, digital twins, inspection, documentation of existing environments, mobile applications and the use of spatial data. This cross-cutting nature is exactly what makes this technology building block particularly strategic for Inoveo3D.

3D reconstruction and its associated processing require considerable computing resources. Inoveo3D will soon be able to rely on a high-performance computing (HPC) infrastructure in Luxembourg, among the most powerful in Europe. This new capacity will let us accelerate our experiments, test different approaches at a larger scale and keep improving the quality of our results. It complements the work already carried out by our team and should allow us to reach a new technological milestone.

Another essential component of our research is naturally artificial intelligence. We are happy to count on the support and resources provided by Google for Startups, which allow us to move forward on the various AI building blocks behind our work. The combination of these environments — our internal R&D, the HPC computing capacity in Luxembourg and the resources made available through Google for Startups — should allow us to considerably accelerate the next phases of the project. We would like to thank the teams and partners who make this new step possible.

This step also carries a particular meaning for Inoveo3D. Since the beginning of this journey, we have chosen to build, experiment with and develop our technologies with our own resources and the funding obtained to support our growth. To date, Inoveo3D has not taken on any external investor in its share capital. We are especially proud to have reached this level of technological maturity while remaining independent. In less than two years, our ambition has gradually evolved: beyond tools for real estate professionals, we are now building technology building blocks around artificial intelligence, imaging and 3D, which can find applications in many sectors.

The road travelled is significant, but we are only just getting started.

This technology is still in research and development. The results presented today are a first demonstration of our approach, and will naturally keep evolving as our work progresses. In the coming months, our goal will be to further improve reconstruction quality, explore new use cases and gradually integrate this technology into our web platform and mobile app. The next stages of our work will be unveiled progressively over the coming months.