Speaker 0 00:00:00 At vision, we changed the way robots, see the world.
Speaker 2 00:00:12 Hello everyone. And welcome to the robot industry podcast. We're glad that you're here and thanks for subscribing I'm Jim Beretta and our guest for this edition of the podcast is Patrice roulette font. He is co-founder and vice president technology after completing his master's degree in image processing and computer graphics, Patrice joined the I N R I a contributing to research in the robotic surgery field. He also co-founded. Intervision where he leads the engineering team. In addition, Patrice contributes to IP growth and the adoption Ofer vision's technology by global corporations. He's overseeing the integration of wide angle imaging technology in numerous industry verticals, and is also an I E senior member invited speaker at various conferences and co-author of multiple publications and patents. So welcome Patrice to the podcast.
Speaker 0 00:01:11 Hello, Jim, thank you very much for giving me this opportunity to, to chat with you today about robotics. So, and especially robot vision.
Speaker 2 00:01:20 Well, and you know what? We were just chatting a little bit, little bit of Montreal because you're based in Montreal and I was actually in Montreal last night and spent most of the night in the airport in Montreal. And so I wanted, uh, you to tell the audience a little bit more about Intervision.
Speaker 0 00:01:35 So in vision is, um, 22 years old company focusing on why Don vision. So we came, uh, back 20 years ago with this idea to create cameras and lenses that mimic humanize. So to enable the, the machine to see wide and in the same time to see in very high resolution. So we combined the best of the world, like as human we see wide, but in the same time, we have an area of interest in the middle of our eyes or our field of view, where we can see very details like, uh, drawing or writing. So that's give us this idea of funding this company, and, uh, to bring eyes to the machines.
Speaker 2 00:02:23 So vision systems for robots.
Speaker 0 00:02:26 So vision system for any type of machines today, we are talking about robots, but in this 20 years journey, we design eyes for surveilance cameras, for cars, automotive, uh, we design eyes going to space for TV broadcast, action cameras, smartphones, tablet, PC, uh, flying drones and, uh, robots. So a long, 20 years drone a designing it for machines.
Speaker 2 00:02:56 What trends are you seeing today? Like, are you spending a lot of time in drone work or is it in Bial robots or
Speaker 0 00:03:04 So thi this is very interesting because we start designing, uh, cameras and eyes for human vision. So back 20 years ago, it was all about human vision. What is the bot the best image to please humans increasing the resolution that in true colors, very vivid colors and sharp image. Now, since five years, we see these ways of AI, what I like to call computer vision and, uh, the, the requirement is changing. So we now still designing cameras for human vision, but more and more for computer vision robots, drones.
Speaker 2 00:03:43 And so, is there any kind of unique research that you're doing in those fields?
Speaker 0 00:03:48 Yes, because, so what is, what make us unique is we combine in the same company in the same lab, which expertise. So first we have lens designer. So we actually designed the lens. So the piece of glass or plastic, one on top of the other that collect the light, uh, in the wall and break that inside the camera. And we have also the camera system, engineers, the image processing engineers, and the machine learning engineers, working all together. And what make us unique is we master the concrete pipeline and we optimize the design of the eyes, according to the result we want to have at the end. And this is a, is a research topic at human vision is what are the best design that maximize the efficiency of the machine learning and AI. So we applied this research to robotics as well as automotive, because for us it's, they, the it's very similar in terms of, uh, visual contacts and vision system requirement.
Speaker 2 00:04:57 So I would call you very vertically integrated, right? Where you have all the, all the right people in the same room at the same time.
Speaker 0 00:05:04 Yes, absolutely. And that's, that's very convenient because instead of working in Zillow, like if you walk with different suppliers, we have this team working together, sit together to find the right solution to the problem we have to solve.
Speaker 2 00:05:19 And what are you seeing as kind of the biggest challenge for, uh, vision systems and robotics and, and machine vision in general?
Speaker 0 00:05:26 So I, I see two main main challenge. So the first one is this combination of, uh, I would say use cases, uh, where we have on one side, the robot vision in the eye. And on the other side, you have the human vision. So it's kind of, you have to design the visual context that please either the robot and the human, and it's not completely the same. So that's bring some complexity to design this kind, uh, visual contexts that fulfill equipment. So that's one challenge that I see in robotic right now. The second one is latency. So it's all about latency. So all you can capture the end environment, process it and extract the right information for the rest of the pipeline. And he has to, he has to be fast because faster is the, the pipeline faster can, the robot can move and take decisions. And, uh, most of the people believe that this, uh, Latin seat is reduced by electronics. Like we had some computing power and more and more computing power, but it start also by the lens as an optical, uh, engineers, optical designers, uh, talk about lens and said, okay, the lens has to be fast, fast, mean capturing lot of light in the small times. So like, uh, that's where we, we bring some contribution is making the pipeline fast, not only the processing, but also the lens.
Speaker 2 00:07:03 And so what does make your approach different? Like obviously using bigger lenses, more glass or more plastic?
Speaker 0 00:07:09 So first we, we are highly specialized in Y donga imaging imaging. So we have 20 years of experience in designing why Doner camera and that's, uh, for me, that's become a mess for robotics. Like you want the robot to capture the food surrounding capture the, the wall. So what makes us different is we are experts in Y Doner camera. Plus we develop this, uh, intellectual properties that allow us to make Y girl camera differently. So you remember the example with human eyes, whereas human. We have Y girl plus augmented resolution in certain areas. So that's what we do considering the application requirement. So let's say we, we build a robot that has to see the world as like human. We design this type of eyes where we see wide with augmented resolution in the middle, but we can have a completely opposite requirement where the robot has to see the surrounding, uh, in eye resolution. So we design different type of eyes. So what make us unique is this ability to, uh, design eyes and vision system with this flexibility of, uh, uh, like we have as a human, uh, with our eyes.
Speaker 2 00:08:24 And so can you give us some use cases of where you're doing that? Cause I know you had a press release just recently.
Speaker 0 00:08:31 Yeah. So, um, the two, maybe most recent, um, uh, use case, uh, I will start by, uh, a project for robotics and flying drones for consumer drones and, uh, the issue they, uh, they have is okay. We can operate outside in bright light, but as soon as we go inside in warehouse or, uh, I would say other low light, dark environment, we cannot operate because the, the camera cannot collect enough light. So we designed those eyes that can, that flying consumer drones to operate in low light in warehouse autonomously. So that's one, uh, one project, uh, we did recently and even more recently, we are teaming up with a human robotic company to design their, uh, vision contacts. And, uh, we are super excited about this project and, uh, this, so we bring this Ultrawide wide vision to their robot.
Speaker 2 00:09:37 Thank you for that. Um, can you elaborate a little bit more on the problem of the dual computer and human vision and what you're seeing as the most viable solution?
Speaker 0 00:09:46 So just to explain what is the problem is, um, so as human, we, we are used to see the world in color first with a certain level of details and we are well designed to, by more natural, right, to, to see the contrast. And, uh, any type of blueness can, uh, make discomfort worse, or even we are very trained to see all the level of green, because there is a lot of green in, in our, and the ment. So we have to pay attention to all those requirements like sharpness color, and that's, uh, required processing and eye resolution. On the other side, machine learning don't care too much about all those details. They can process, uh, very, um, low resolution image and take decision out of it. And the, I would say the main focus is latency. So all we can go fast, uh, with, with AI. So it's kind of competitive requirement where we have to build a system for human vision and competitor vision slash AI in the same device. And the way we solve that is we add different layer of image processing, depending if we serve pixel for human, or if we serve pixel for machine learning and AI.
Speaker 2 00:11:09 Oh, and that's very clever. So how do you work with your clients? So I'm wondering if you're doing a lot of education with your clients because you you're the industry experts, right?
Speaker 0 00:11:19 So we have, uh, so first we enjoy very much to team up with our clients. Uh, we feel that, uh, they know very well what they do. And on our side, we know very well, the imaging, why imaging visual contacts. So what we like is we, we like to team up and join their teams, uh, to help them to solve the, the, the, their problems and to do so. We created, um, a team in team vision. We call that innovation lab. Mm-hmm <affirmative>. Uh, so the beauty of this team is there are experts in vision system, plus data. They are working in different industries. So they are working in automotive via space, uh, broadcast, consumer electronics. So they have a broad spectrum of, uh, I would say project and knowledge that they can bring to a specific industry like robotic today to innovate and solve a customer problems.
Speaker 2 00:12:22 And Montreal is kind of a, a big place right now for vision, for robotics, for, uh, machine learning for AI. So does that kind of help your business a bit?
Speaker 0 00:12:33 Uh, very much because yes, you're right. So Quebec Montreal is, uh, is very well known for AI computer vision. And, uh, that just to add on top of that is, is not only today, right? It's the more Montreal as like 20 or 30 years is still be in with they Montreal host great company in imaging. And so that's that motivates us to move vision from France to Quebec. Uh, I would say 17 years ago, we are very lucky to have this ecosystem of, of companies and university in Montreal and Quebec, uh, to help us to sell our work customers.
Speaker 2 00:13:17 Oh, that's great. And Patrice, what kind of work are you excited to do now? And because you're doing so many different things.
Speaker 0 00:13:24 So the, since we're talking about robotic, uh, so all these work combining, uh, VI vision system, visual CORs, and machine learning is super exciting for us because we, we see that as I would say, transition between human vision and computer vision and designing system for this specific application. It's, uh, it's super exciting on the side of that. Uh, we do a lot of work also on tablets and PCs, uh, TWIs to increase the, the, the quality of the camera inside those device. And also is the user experience of collaboration, like what we are doing today, uh, true, uh, true video conference. So we, we are pleased to have this opportunity to contribute in this area mm-hmm <affirmative>. And in the same time we prepare, we are preparing the future. So there is a new trend in objects to make the cameras smaller, more efficient with more functionalities. Uh, we call that flattop optics or Netherland, and, uh, immersion is, is doing lot of research in this area of flat optics and Netherland to prepare the future of the camera.
Speaker 2 00:14:37 So Patric people must be very excited to come and work for you.
Speaker 0 00:14:40 Yeah. So, so all we all, we try our best to attract talents and, um, we decide 15 years ago to, to build those talents. So let me explain. Uh, so we invest in university. We are the sponsor of a researcher in, uh, LA university in Quebec. And we are proud to say that we are almost all the, the lens designer coming out of this program. Uh, we also tighten our collaboration with computer vision and AI, uh, university, uh, in Quebec as well, and in France. Uh, so again, we hire, we, uh, we scoot hire the, the good talent, uh, plus, um, I would say that we are in, even if we are based in Quebec, we are, we have operation worldwide. So we are looking for talents everywhere in the world. Uh, and, uh, this new weight work remotely help us to, to have people working for us, even from, uh, in those part of the world.
Speaker 2 00:15:52 Yes. It's an exciting time to be in the automation, robot, vision optics, uh, industries, of course. So when, uh, when you're not designing new optics and helping customers out, what do you like to do?
Speaker 0 00:16:03 So, uh, I love surfing. Uh, so you say that, oh, you would say that Quebec is not a place for surfing <laugh>, but it's not true. Uh, we have a, a static wage in the Lawrence river, so when I'm not traveling for, I can in Quebec, in Montreal. So that's one of, uh, that's, that's one of my Obi, the deal, the Obi is, uh, I like connection. So, uh, I like to connect with people. And, uh, so each time I have the chance to travel for conference or meet customers. I like to, to connect with people in the industry and catch up on what is new, what we can do better, differently, all we can innovate. So that's, that's also the deal Myk I would say that, uh, uh, I, I do it and I enjoy very, I enjoy it very much.
Speaker 2 00:17:02 Well, um, thank you very much for, uh, coming out to the podcast and telling us all a little bit about Intervision and how can people get a hold of you if they'd like to find out more.
Speaker 0 00:17:11 So first, we, they, they can start by our website. Uh, so there is different information and especially we are about to really the section dedicated to robotics. So they will find some information there. They, they can contact us through, through the website. And, uh, we, there is always somebody, uh, expert in the field of our customers to answer to that question connect, and then, uh, we see how we can help.
Speaker 2 00:17:42 And tha thanks for spending a bit of time with us today,
Speaker 0 00:17:46 Jim, I, I would like to thank you for this opportunity. I really appreciate, uh, this time and, uh, the opportunity you gave me, uh, to, to explain, uh, what immersion is doing. And, uh, uh, and especially on my side, personally, why I'm doing that is I feel that since I, I was young, I, I Al I always had the passion for vision. So understanding the vision and the lack of sight. So if we lost the vision, so that's motivate me to, to work in camera image, processing vision, and I enjoyed that very much. So I, to share that with you today,
Speaker 2 00:18:25 Thank you. Our, uh, sponsor for this episode is Airhart automation systems Airhart builds and commissions, turnkey automation solutions for their worldwide clients. With over 80 years of precision manufacturing, they understand the complex world of robotics, automated manufacturing, and project management, delivering world class custom automation on time and on budget contact one of their sales engineers to see what Airhart can build for you. And Airhart is spelled E H R H a R D T. I'd like to thank and acknowledge a three, the association for advancing automation. They're the leading automation trade association for robotics, vision and imaging motion control and motors, and the industrial artificial intelligence technologies visit automate.org to learn more. And I'd like to thank painted robot. They build and integrate digital solutions. They're a web development firm that offers SEO and digital social marketing, and can set up and connect CRM and other E R P tools to unify marketing sales and ops. And you can find
[email protected]. And if you'd like to get in touch with us at the robot industry podcast, you can find me Jim Beretta on LinkedIn. We'll see you next time. Thanks for listening. Be safe out there. Today's podcast was produced by customer traction, industrial marketing, and I'd like to recognize my nephew, Chris gray for music, Jeffrey for audio production, my partner, Janet, and our partners, a three painted robot and our sponsor Airhart automation systems.
