In this episode of Visions: A Machine Vision and Automation Solutions Podcast, VSD's Sharon Spielman interviews Hantz Fevry, CEO of Geolava, about spatial intelligence — AI that understands and reasons about the physical world in 3D — who explains how world models enable vision systems to perceive, predict, and make smarter decisions across time and sensors.

Fevry discusses key enabling technologies (multimodal imagery, LIDAR, temporal fusion), practical factory-floor applications (co-pilots, predictive maintenance, robots), integration challenges with legacy systems, validation advice, and how leaders should focus on measurable outcomes and quality multimodal data.

Related: Embedded AI and Smart Cameras: The Next Decade of Machine Vision

Visions: A Machine Vision and Automation Solutions Podcast, is the podcast for engineers, designers, integrators, and end users who want to keep an informed eye on the imaging and machine vision industry. Every Tuesday we will explore the latest in imaging trends, developments and solutions. Here you will find interesting, useful insights and observations from expert interviews, solo episodes, even the occasional panel discussion, all of which aim to expand your knowledge on imaging and machine vision. 

Transcript

Well, hello and welcome to “Visions: A Machine Vision and Automation Solutions Podcast.” I'm your host, Jim Tatum, senior editor of Vision Systems Design and Visions is an Endeavor Business Media production from your friends at Vision Systems Design. Here you'll find the latest on everything from end user machine vision solutions to trends, developments, and perspectives on all things machine vision and imaging. Whether you've been working in the industry for a while or you're just starting to take a closer look at it, this podcast is designed to grow your knowledge and bring greater focus to your understanding of the imaging and machine vision industry. And now on to our show. 
Well, hey everybody, and welcome to Visions. Here's a question for you. What if machine vision could do more than detect objects? What if it could understand the world around it, predict what happens next, and help machines make smarter decisions in real time? Well, as it turns out, that's exactly what spatial intelligence, a rapidly emerging AI capability does. And it's changing how vision systems, robots, and automation platforms interact with the physical world. Recently, Vision Systems Design’s head of content Sharon Spielman got to spend a few minutes with Hans Fevery, CEO of Geolava, a San Francisco-based company that specializes in spatial AI. For a firsthand look at how spatial intelligence is reshaping industrial imaging system integration and the future of AI on the factory floor. Hello everyone. Today we are focusing on spatial intelligence, the AI capability that lets vision systems understand and reason about the physical world in 3D. This breakthrough is influencing the future of industrial imaging, robotics, and automation. Joining us today is the CEO of Geo Larva to share his insights on how spatial intelligence is transforming vision, system design, integration, and practical deployment on the factory floor. Thank you so much for being here today, Hans. 
Yeah. Thank you very much for having me. 
So I will just jump right in. Um, in simple terms, what is spatial intelligence for machine vision and why is it getting so much attention right now? 
Yeah, that's a great question. So spatial intelligence enables AI to perceive, understand and predict the physical world across space and time. So it doesn't just recognize images. It represents the next evolution of AI moving from just the perception into meaning. So, basically the reasons, uh, with the physical world. So if you think about AI systems today, you have one digital system talking to another digital system right now with spatial intelligence powered with world models, you're able to bring AI to the physical world. One key example is that a camera can see a crack on the floor. A spatial intelligence platform powered by World Model will understand why it was formed, how it's evolving and whether it will become a structural risk or not for a property. 
Okay. Thank you for that. So how does spatial intelligence enhance vision systems ability to see and interpret complex industrial scenes? 
Yeah. So, uh, of course, it gives machines the understanding of the state, uh, the relationships and the evolution of the physical environment that will allow them to basically reason about the entire scene instead of just detecting objects in isolation. So instead of detecting, for example, a forklift and a worker separately, it will understand there is a collision path. And because of the collision path, there is something at risk. So there is a bit more of the reasoning beyond just the perception or the object detection. 
Okay, so what are the key technologies that are enabling the spatial understanding, especially in terms of vision? 
Yeah. So it's a multi-modal platform. So a world model will focus a lot on the perception. So leveraging different type of imagery. It can be optical LiDAR, thermal uh, 3D scene. And then as well as a contextual layer, it can also bring some text data and then basically understand not only what it's seeing, but the evolution of what it's able to see. So this is how you'll be able to build those spatial intelligence that just like a human, will be able to interact with the physical environment. 
Okay. Uh, so what have you seen to be the biggest practical challenges when integrating spatially intelligent vision systems, in particular manufacturing? 
Yes. So the challenge will be it won't be only collecting data. So it will be creating a reliable model that can continuously understand dynamic environment, um, that can adapt to change and integrate with existing industrial workflows. So that most of the time our legacy systems. So, um, that is a bit of a challenge where you have an industry that works with a lot of very solid workflows or systems. And then you have that frontier AI model that can perceive that can reason, that can understand and predict. And how do you make sure that the legacy system can work in tandem with those new technologies? 
Okay. How does this spatial approach improve the flexibility and the accuracy over 2D vision systems? 
Yeah. Um, it improves it tremendously. So traditional system in 2D will recognize what's visible in a single image. So it will recognize an object. It will try to make some measurement depending on the type of imagery you're using. With a spatial intelligence, you can really understand the physical world itself, making it more robust, adaptable and capable of anticipating future states. And basically, what you can do is that you not only understand the relationship or the causal effects of a lot of features in the present, but also can forecast into the future a future. So you go way beyond detecting objects. 
All right. Can you talk a little bit about the role that AI driven image processing algorithms play? 
Yeah. So the AI will they transform raw observations into a structured representation of the physical world. So they will they can provide that foundation for reasoning for forecasting or prediction and decision making. So with AI, you can be way more informed to understand not only what you see, but how it evolved into that current state and how it will continue to evolve into future state and forming better decisions. 
Okay. How can the spatial intelligence help vision systems better handle the tricky issues such as occlusions, reflections, inconsistent lighting? 
Yes. So when you focus a lot on one image or set of images, then this is where you're going to have those type of challenges. But with spatial intelligence or world model, instead of relying on a single image, you can focus a lot on fusing information across sensors, viewpoints, time to infer, and what basically cannot be observed, because the world model will be able to reconstruct beyond what it sees from the from the pixel. So in that case, those type of challenges are no longer present because of that reasoning aspect. It is similar to a human being. Sometimes the human being doesn't have to see a place or a room with perfect lighting. To be able to make an assessment, you will be able to understand or even forecast what it should be, what should be there even without that perfect vision or clarity of the place. Okay, so it's not it? Well, maybe it is. It's using AI to, um, replace what it thinks the lighting should be or replaces what the reflections are. No. So basically what will happen typically on world model is that if the lighting is not perfect there. So think of it as an image and there is a lot of mask, uh, on top of that image, it will be able to predict what should be there, uh, even without being able to see it perfectly because you understand causal causality and relationships. So even if there is not a perfect environment, perfect lighting, then with that spatial intelligence, it can still navigate and see through it. But beyond that, um, the spatial intelligence, what it focuses on and what it's good at is basically it's on the reasoning. So it depends on what you're trying to accomplish. If you're trying to make an observation of a place to detect a risk of a stroke or I don't know, any structural risk in that case, without that perfect lighting, you can still have a level of understanding of that risk be instead of focusing on a single image, that will give you a binary answer. 
Okay. What about on the factory floor? More of our readers are into industrial manufacturing. So they're looking for solutions on the factory floor, you know, with the machine vision that they're using in, say, their inspection things on the manufacturing line. 
So, um, Steve Anderson, correct the, the question correctly is what is the application for someone in the factory floor leveraging that compared to legacy system or legacy computer vision systems? So I would say right now on the factory floor, you can have a true copilot that not only help you understand and make faster analysis, but make analysis for you and then help inform your final decisions. So then the person in the factory floor becomes way more efficient. And that spatial intelligence can be powered, uh, powering, uh, either a robot. So this is where we're talking a lot about physical AI, or it can be powered, just, uh, powering just a software and I help make an analysis and help make some forecasting decisions. So this is some of the many applications that can be used when we bring world models to the factory floor. 
Okay, great. So what advice would you give engineers and integrators for validating and benchmarking these advanced vision systems? 
Yes. So this is where hallucination will come in. And I think what's very important is to benchmark real world understanding, not just detection accuracy. So it's important to measure how well the system generalizes reasonable change and predicts outcomes in an operational environment. And then always measure against those key benchmarks and which is not easy but also not impossible to do so. The real test is not whether it detects a defect today, but whether it predicts the defect before it becomes a failure. And what is that benchmark? What is that ground truth of a defect. So that will be basically um the advice I would share for any engineer trying to implement any world model or spatial intelligence system to their workflow. 
So from your perspective, how will spatial intelligence change the future of vision hardware and software design? 
Yes. So right now, I think vision system or passive perception that detect objects that basically make sometimes some connections, but right now that those vision systems will evolve from passive perception tools into world model that continuously perceive, understand and simulate the physical world. So just as language models became the intelligence layer for text, world model will become the intelligence layer for the physical environments. 
Okay. So what should manufacturing leaders focus on to build confidence and readiness for adopting these new spatially intelligent vision technologies? 
That's a great question. I would say to start with measurable business outcomes. Uh, to invest in high-quality multimodal data and prioritize platforms that continuously learn and improve as the physical world changes. So for example, so begin with a production line where reducing downtime by ten percent delivers an immediate and measurable return on investment rather than just testing the spatial intelligence for the sake of having a new technology there, because sometimes some legacy systems are robust and work, and the spatial intelligence or the immune system can only be useful in a very, very specific use cases. So basically identify those use cases and implement those spatial intelligence or world model there. 
Okay. Uh, do you want to tell me a little bit about Geolava and how long it's been around, how long you've been the CEO? If you could tell us about that. 
Sure. So I've been working in AI for the past ten years. I was at, uh, working at Google at part of speech that is now part of Google Deep Mind. So I work a lot on deep learning. So after Google, I built a company that we sold in twenty 2024 and now we're building Geolava and what we're doing is we're bringing AI to the physical world. We're building a world model for, uh, critical infrastructure and properties so people can not only understand their current state, but forecast their future state. So that's what we've been working on. And our team from DeepMind, Apple, Meta with very strong experience in computer vision, as you mentioned, and as well AI in those frontier models. 
Okay, great. Is there anything else you want to tell our audience? They are the engineers and the engineering managers at OEMs, uh, with machine vision and image processing. Uh, so what would you like to say to them? 
I would like to say to, um, definitely make a lot of research on world models and not to confuse VMs or LM for world model because they're very, very different. And yes, to, as I mentioned before, to try to see where a world model or a spatial intelligence can bring value and, and measurable ways. And yeah, in my opinion, world model will revolutionize the AI and the physical world the same way. LM revolutionize, uh, AI with Dex. 
Great. Well, thank you so much for joining us today and providing this information. It's going to be a really exciting time in our industry as we move into this physical AI space. So I really appreciate you speaking with us today. 
Thank you very much. I appreciate the time. 
Well, that's a wrap for this episode of visions produced by Endeavor Business Media, a division of endeavor B2B. Thanks very much for tuning in. If you enjoyed today's show, be sure to subscribe to the podcast and share this episode with a colleague who would find it helpful. Until our next episode, you can find us at vision dash systems dot com or on LinkedIn, Facebook, or for more insights, updates, and breaking news to keep you in the know. Thanks for tuning in. Until next time, stay focused on your visions.