When the camera thinks for itself

Dr Basel Fardi & Dr Heiko Cramer

The start-up company Intenta develops software and algorithms for driver assistance systems as well as intelligent, camera-based sensors for automatic scene analysis. The company, which was founded by two former doctoral students at Chemnitz University of Technology, won the prestigious "Macher 25 - the big business prize of the East" award in the "Start-up/Innovation" category and hired its 100th employee at the end of last year. That's why the two founders Dr Heiko Cramer and Dr Basel Fardi are the movers and shakers of the week for us.

Founded in 2011, your company was honoured last year with the "Macher 25 - der große Wirtschaftspreis des Ostens". "Where it says Audi on it, it often says Intenta in it" was the reason given. Does that aptly describe your activities?
Dr Heiko Cramer: There are indeed cars that contain our software, but "often" is a bit of an exaggeration. Our great success is that such a young, still small company manages to get software components into production vehicles from major manufacturers.

Your software can be found not only in driver assistance systems, but also in the SmartSensor S2000 developed in-house - where else?
Dr Basel Fardi: The company was originally founded with the idea of an innovative, intelligent sensor for the civilian surveillance of people and objects: A camera that thinks for itself, or rather transmits information to the user, thereby making surveillance more efficient. This idea has since developed into a product ready for series production, which is gradually being introduced to various markets. One of the first areas of application is currently hospitals. In radiotherapy programmes, we monitor the treatment room and ensure that the radiotherapy process is only started when there is no one else in the room apart from the patient. In this way, our sensors prevent the unwanted irradiation of staff.
At the same time, we have always had our eye on the automotive sector and have developed components for driver assistance systems. Right from the start, we were able to win the Volkswagen Group with Audi and Co as customers. As a result, the automotive sector grew very quickly and the company initially grew up with it. However, intelligent monitoring and thus the area of sensor technology is also becoming increasingly important

How does the technology in your SmartSensor work?
Dr Basel Fardi: Until now, the cameras were "stupid", if you like, and only provided two-dimensional images. An employee sat in front of countless monitors, observed them and reacted when a certain event occurred. The problem with this is that an incredible amount of image information is transmitted simultaneously. Added to this is the laziness of the human eye; the observer becomes tired and unfocussed over time. This is why monitoring is not really efficient enough. The ideal solution is a camera that thinks for itself, so to speak, and alerts the observer to events in good time.
The principle: humans are created with two eyes and one brain - the eyes see the images and the brain processes them. This enables us to process the environment three-dimensionally, i.e. spatially. We imitate this principle with the sensor. It has two cameras (the "eyes") and a processing unit in the form of a brain that interprets the images. Our expertise lies in developing the intelligence, the "thinking" of the camera, so to speak. In short, we teach the camera what objects, rooms or people look like and which behaviour patterns are normal or conspicuous.
We are currently testing an application in the field of AAL (age-appropriate assistance systems for a self-determined life) in collaboration with GGG. The aim is to recognise emergency situations, such as falls or inactivity, in the home environment in good time and automatically and to organise help by raising the alarm. In terms of data protection, it is important that the image information is processed directly on the sensor and that the images do not leave the camera, as this would infringe on the personal rights of the individual. Therefore, only the meta information about the event is output - e.g. the patient has fallen.
Another possible application is in areas with high public traffic, where crowds of visitors are counted (in department stores) and controlled (at airports). The system could recognise long queues and trigger the opening of an additional check-in counter.
Back to the car: the next big step is certainly autonomous driving, and it's not just Google and Tesla that are researching this. How long do you estimate we will have to wait before the first self-driving vehicles are on German roads?

Dr Heiko Cramer: The first vehicles will be on German roads before 2020, that much can already be said. However, they will initially operate in very limited areas. The initial plan is to implement semi-autonomous driving in traffic jams on motorways and then work our way from there via rural roads into urban areas. The more you move towards the city, the more complex the scenarios become and the more difficult it is to actually have a car drive itself.
This initially works best in well-structured environments such as motorways. There are practically only white lines there, there are usually no people and oncoming traffic is separated by crash barriers. These are all important factors for autonomous driving.
It will certainly be some time before it works in the city. Google is already testing with a special licence in California. Technically speaking, a car could already drive autonomously today; remote control via computer is not the problem. The challenge is to make it so safe that the driver can consciously focus his attention on something else. It is also an important legal question as to who then assumes responsibility. As far as we know, this is still unresolved in Europe. But technically, we are actually very close to addressing this issue by summarising the data collected by the numerous sensors in the car and thus describing the vehicle's environment fairly accurately: Speeds, static and moving objects in the environment, turning lanes, etc. All this information is then used to derive appropriate recommendations for action.
Incidentally, we have a special licence for Germany that allows us to build test vehicles ourselves and use them to test our software on local roads.

You met at university in Hamburg and then came to Chemnitz to do your doctorate - and have described this as a culture shock. What did you mean by that?
Dr Heiko Cramer: I knew that it would catch up with us at some point ... (laughs).
Fardi: Well, we came to Chemnitz from the cosmopolitan city of Hamburg in 2000. The cosmopolitan attitude that prevails there wasn't immediately apparent here ... (smiles) But in the 15 years we've been here, Chemnitz has developed massively, you just have to say that. Just comparing the city centre from back then with today, it's a shame that we didn't take any pictures back then. A lot has happened in the meantime and that's why we've stayed here. If you have a family, a nursery place here is simply a matter of course. But many people overlook these advantages.

... and that's another reason why you've remained loyal to Chemnitz to this day?
Dr Heiko Cramer: Let's put it this way: it was never an option to set up a company elsewhere. I don't know if we would have stayed here if we hadn't founded the company. We would probably have gone our separate ways and returned home. I am originally Swabian. But when we decided to set up Intenta, it was already clear "if we were going to do it, then here". Because the region has a lot to offer. Not to mention the good conditions and support for young companies here at the TCC, the funding opportunities in Saxony - it was a perfect fit.

The company has now grown so much that it is too big for the TCC. Intenta now has over 100 employees. The much-vaunted shortage of skilled labour doesn't seem to be an issue for you. Is it because of the proximity to the TU?
Dr Basel Fardi: That's right. We both come from the university and still have a direct line there. We know a lot of professors and staff. It is extremely important that we have the TU close by. That's why we are also involved there in the long term through an endowed professorship. But we have also been able to attract new colleagues from the surrounding area. Good people regularly come from the universities in Mittweida and Zwickau. And from time to time, the odd returnee comes to us.
But of course we also realise that there is a shortage of supply. During the start-up phase, we were initially ridiculed when we were the first to set up an HR department. But we knew from the outset that the need for skilled labour could become a problem, which is why we paid particular attention to the issue.
Dr Heiko Cramer: Our employees appreciate working in a relatively young team and a relaxed but professional environment. What's more, tinkering with cars that usually have a lot of horsepower under the bonnet or test drives such as those currently being carried out in Iceland or Malta is naturally exciting for young engineers.

You are currently planning to move to a location steeped in tradition, the former Union machine factory on Ulmenstrasse. What tipped the scales in favour of this location?
Dr Heiko Cramer: First of all, we are building a modern, new company headquarters next to it and will only be renovating part of the historic Union hall. The building is really quite large and not so easy to modernise. That's why we're moving into the new building first and will then renovate the old hall bit by bit and take it over.
The location is of course great: it is close to the city centre, most of our employees live close to the centre, many on the Kaßberg. That's why there was never any question for us of building somewhere far out on a greenfield site. The area is very well connected by public transport and it's easy to get to work by bike. The area is large enough, so we have enough room for manoeuvre for the future. And in the end, we split it up a bit: I like historic industrial buildings, my colleague prefers new buildings. We both have fun with that ... (laughs)

Last question: Do you have to encourage the people of Chemnitz?
Dr Heiko Cramer: I don't think you have to encourage the people of Chemnitz. But in my opinion, they are really selling themselves short. Encourage is perhaps the wrong word. You should say that they can be more self-confident - even to the outside world. "Encourage" sounds like it's terrible here and you have to persuade them to stay. But I don't see it that way at all!
Dr Basel Fardi: and when it comes to encouraging young people to set up their own business: As I said, the region has a lot to offer. Fortunately, the West/East issue is playing less and less of a role. That's why the locations here are becoming more and more attractive. In this respect, young people should definitely be encouraged to think in the direction of self-employment. Of course, you need good ideas, but Chemnitz has a great history in engineering that you can build on.