How Tinder helped us to create a smarter LPR system

Januar 16, 2018

Musings of the CEO

Ok guys, I’m totally aware that the title is as pure a clickbait as it gets – but let’s be sincere, you are here reading this piece because of that very title. And I also promise that you will soon understand the connection between Tinder and real smart LPR if you continue reading.

So, let’s find out where the link between dating on an online platform and good LPR is. What happens on Tinder is that you get a picture of a possible love interest and get to decide to sweep right (like) or left (dislike) and see if you match. Let’s imagine you sweep right and match. Does that mean that you are marrying the subject the next day? Of course not. With all due respect to those who think otherwise, nobody in their right minds would marry anyone based on one picture. You start chatting, dating and get to know each other during a series of social interactions before getting together for life. This much is obvious for human relationships of the romantic type but there are other fields where the same pattern could and should be applied. Let’s take the LPR market; for some reason, we generally do not use such validation techniques and literally marry based on random Tinder pictures.

Just think it over. The LPR system receives a sole picture after some sort of process followed by applying the business logic of the system. All based on a sole picture. Is that really all right? I guess not. Otherwise – sticking with the Tinder example – this would mean an increased workload for divorce proceedings lawyers. This phenomenon on the LPR market is called false positive or no recognition. False positives are especially tricky as for instance in case of parking applications, the recognition rate is top notch. The only problem is when you get a “Christian” instead of the “Christina” you were expecting. Definitely not something to be happy about.

This is why we at Asura Technologies believe based on a video stream instead of requesting picture after picture until an event can be validated. Video stream based validation is much more sophisticated – it is like going on multiple dates with our love interest in a really short span of time. ARU has a much clearer understanding of the information stored in the picture thanks to a continuous “correspondence” between the engine and a multitude of pictures broken down to the level of pixels. Still, why use a stream and hunt for information amidst the change in pixels instead of using pictures? Easy. It may happen that you like someone looking from the front, but taking a closer look at the back is a deal breaker. Sorry for the metaphor.

ARU has no need for physical triggers, it starts capturing information on the license plate of the passing vehicle early on and this is why we get to piece together the perfect picture from multiple snapshots taken at different points of time. We know exactly the license plate number by the time the vehicle reaches a barrier. This is how we – following the Tinder example – make sure that you marry the woman of your dreams and even confirm that her family is tolerable.

ARU delivers maximal recognition rate while keeping false positives at a minimum. Oh, and did I mention that the system can be connected to any kind of IP camera system – plug play, no need for time costly integration – able to recognize any license plate in the world.

So how about a date on Intertraffic in Amsterdam? We will be there – with more than a sole picture.s

Click Here to Download our Demo