craftar
Catchoom Blog

Catchoom CEO on Future of Image Recognition and Augmented Reality in AWE 2016 Video Interview

icon Date Jun 13, 2016

icon author Zsofia Kerekes

icon folder , , , , , , , ,

After being interviewed at AWE 2015, we returned to AWE TV’s black table for a hearty chat about the future of Catchoom and the augmented reality and image recognition industry in general.

Many thanks to Mark Piszczor ‒ who works at Occipital when he’s not performing voluntary news reporter duties for AWE media ‒ for giving us the opportunity to share our insights and some big announcements.

Check out the video interview or read the transcript to learn more about our new industrial 3D object recognition technology, our vision of getting beyond markers and depth sensors, our new visual similarity research project for Fashion and a very special 13th anniversary.  

Read the transcript of AWE TV’s interview with Catchoom’s CEO, David Marimon:

AWE TV: Welcome to AWE TV, my name is Mark Piszczor and with me is the CEO and Co-founder of Catchoom, David Marimon.

David, you are an alumni of AWE TV. We interviewed you last year where you made some predictions. I asked everyone what were their predictions about what we’ll see this year. Sounds like these happened; you were right.

Something Catchoom has been working on is 3D object recognition. Let’s start with that. Tell me more about what Catchoom is doing in that area.

David Marimon, Catchoom (DM): We try to stick to our promises. Last year I was saying that we will be working on industrial applications and finding solutions for the industry.

This year at the show, we’re talking a lot about 3D object recognition. Basically, it means being able to point a device like a tablet at some part of an engine and being able to recognize it, based on the structure of the object.

We use a depth-sensor for that. We use the Occipital camera and our own computer vision on top of that. We’re actually seeing a lot of interest based on this demonstration.

AWE TV: You’re providing more than just a 2D marker to recognize an object and its individual components. How do people feel about this ‘3D’ object recognition?

DM: We’re attacking this problem quite differently from others. WMarkerless-3D-Object-Recognition-Catchoom-1.pnge don’t want to stick markers on engines. We want the object to speak by itself.

It can work as a trigger for augmentation and then, for instance, offer guidance to field operators. From maintenance to training applications, it’s a necessary step that serves as the beginning, the ‘trigger’ of an experience.

When I talk about the interest people have been showing ‒ for example when they are reaching out to us after we send them a teaser video ‒ I mean that they want it and they see that it is a necessary step for their own platforms or for their own experiences.

AWE TV: What are they saying about it once they actually saw it happen. Do they say “this is great, we want more”? Tell me about that experience.

DM: They have a bunch of questions. These range from “What are the limits of the technology?” to “How easy is it to embed in our own platform?”.

Depth-sensors have a number of limitations, for example, not every surface has the same reflectance [see reflective surfaces, influence of light etc.]. So, many questions refer to the limits of the technology.

Structure-Sensor-Mobile-Object-Recognition-Industrial-Catchoom.jpg

We’re happy to say that we’ve built some amount of computer vision on top of the sensor to be as agnostic in terms of the device as possible.

Actually, we’re also working on more computer vision to be independent of the camera, reaching a point when we won’t need any depth-sensor.

AWE TV: You mean, independent of the sensor, not the camera, right? You would still be utilizing the camera.

DM: Yes, we’d still be using an RGB camera. The challenge here is whether we can completely remove the depth-sensor and use only computer vision. We’re working on that.

As I said, other questions include how easy it is to integrate it into existing platforms. Right now we are demonstrating it on an iPad, using iOS. However, our algorithms have always tried to be as horizontal and platform agnostic as possible. So, we’re confident that we can integrate it into different platforms.

AWE TV: Something I’ve also heard is that people want more ubiquity across platforms, they want platform agnostic. What are the limitations of one platform over another? Are companies using Android over iOS?

I’m hearing a little bit about Windows coming up, too. Windows 10 is kind of a deal breaker as companies are already using Windows 10 and Windows operating systems. Now they are bringing it to the mobile devices that the company uses. What’s the spread that you’re seeing?

DM: I think the question is not specific of an operating system as much as it is about being able to run completely on device or whether you need cloud support.

Ours can run completely on device. In this industry environmenscreen-shot-5.pngt, it’s important, because in some applications, you cannot rely on the connectivity. You may even have to remove the WiFi and Bluetooth for security reasons, for instance, in defense or aerospace projects.

AWE TV: In some scenarios, there is no WiFi connection…

DM: Exactly. So many questions revolve around that. They also often ask which operating systems we support. At the moment, we developed this on iOS, but we are confident that we can deliver on other platforms.

AWE TV: Very cool. Did other things that you mentioned last year come true this year?

DM: Last year, I said that we would be recognizing clothes, such as garments and accessories.

The answer to your question is ‘yes’, as we’re also delivering a sFashion-Visual-Similarity-DeepLearning-Catchoom.pngolution that is especially meant for e-commerce. It’s not so much about augmented reality as it’s about being able to recognize garments to facilitate a purchase online.

That said, you can provide recommendations based on visual aspects.

Instead of saying “As you are interested in this, check out these other items customers bought”, you can say: “As you are interested in this, I can show you other items from the catalog that look very similar.” [For example, based on the pattern.]

AWE TV: That’s a need I haven’t considered. I am a big fan of AR, so in my mind, I’m trying to understand how that fits into AR. But you’re saying that it’s not necessarily an AR application.

DM: Correct. At Catchoom, we’ve been developing in the space of AR, as well as image recognition and object recognition, and we’re continuing that. As for me in particular, I’ve been working for quite some time in AR…

AWE TV: Actually, today marks 13 years of you working in AR.

DM: Yes. 13 years ago, I started with my PhD. This was when I was in Switzerland and since then I’ve been working in AR. At the beginning I was using ARToolkit, back in the day when it was only through fiducial markers. I worked in SLAM, I worked in tracking and image recognition.

AWE TV: You’ve seen it progress.

DM: Yes, of course.

AWE TV: Sorry to interrupt you talking about Fashion and Catchoom in this area.

DM: We are quite broad in what we offer. I think augmented reality has to do a lot with rendering content. The underlying technology, if it uses tracking, requires computer vision.

Catchoom offers computer vision for a bunch of things. One of these is delivering tracking for augmented reality.

AWE TV: I got you. Very cool. Last year, I asked you about what to expect this year. What can we expect Catchoom to provide next year, or what do you want to see at AWE next year? Where do you see the industry heading in the next year?

DM: We have a significant strategic bet on these two things that I mentioned last year, Fashion and Industrial applications.

I think that in one year’s time, instead of it being something new to us, it will be a more mature technology that we’ll be offering, including solutions in this space.

My bet is that by then, we will also have a stronger relation with other companies here at the show, integrating our platforms and offering solutions for the industry.

AWE TV: I hope so. One of the reasons why I keep bringing you back on AWE TV is that Catchoom has been around for some time. You are also experienced ‒ we’re talking about 13 years. You’re not new to AR and it’s great to have you back on AWE TV.

And I think you’re right. I think that a lot of AR tech offerings and services have to mature and we’ll see a bigger AWE next year, just as it is bigger this year from last year.

David, thank you very much.

If you have any questions or you are interested in our vision and solutions, we can’t wait to hear from you:

Talk to our team

The interview was partially edited by Catchoom to aid readability.

Recent Posts