Image recognition app: Did you ever think you’d see a fish drive by? You saw it right. Image recognition knows no bounds.
It’s been a while now (2014) since “Fish on wheels” went viral. Did you miss it? No problem, we’ll tell you again. The Fish on wheels, (FOW) is a self-steering fish bowl on wheels. The FOW consists of a chassis of a steerable car and a water tank with a structure consisting of a smartphone with an image recognition application above it. The smartphone’s camera records the position of the fish based on a grid of nine planes. Does the fish swim in the right-middle plane? Then the car incl. the fish drives to the right. Does the fish swim in the middle-left plane? Then the car drives to the left and so on. At the time, the FOW was a project to explore the possibilities of image recognition. Despite this, the story of FOW went viral and BBC, among others, covered it.
Image recognition application and 5G
The FOW works through an image recognition application and 5G. But what does a moving fish have to do with image recognition and 5G? Don’t worry, it’s not a conspiracy theory or rocket-science, you can leave the tinfoil hat in the drawer.
The human eye is able to see its surroundings, just like a camera. But the brain is needed to link meaning. Like recognizing people, looking for stuff or recognizing the color of your own car. In this case, the Cloud is the brain behind the camera. The camera is the proverbial eye.
The camera records data that can then be linked to the execution of an action. In the case of FOW, an image recognition system recognizes the orange color of the fish, and based on the position of the color orange (the fish) links an action that is performed by hardware. In this case, that is driving and steering a motorized set of wheels. The main element of image recognition is the camera, but the data recorded by the camera goes directly to a server in the 5G network via a 5G connection. This server converts this data to an action to be performed and sends it back to the smartphone which ensures that this command is executed by the hardware. This allows live images of photos and videos to be processed in real time and with minimal delay.
Why 5G and not just a smartphone application? You may be wondering. Using 5G and a separate server avoids heavy applications that consume a lot of data on mobile devices. All the complicated and heavy data sets and calculations are done in the Cloud. This makes deploying image recognition on mobile devices easier and more accessible.
Making the most of 5G speed
Now that smartphones have a lot of computing power and the connections are so good due to 5G networks, among other things, two options have emerged for implementing image recognition. It is possible to set up an app so that image recognition is performed on the phone itself, or to upload the image to a server where it can be processed and the result sent back. Both ways have their advantages and disadvantages.
On the smartphone, computing power is somewhat more limited, but the user is not dependent on a connection and its speed. However, an improvement to image recognition will have to be made by an app update or updating the data in the app again. In the Cloud, on the other hand, computing power is abundant and an update can be rolled out to millions of users within seconds. The downside, however, is that data always needs to be sent back and forth.
In addition to riding fish, 5G and image recognition a combination that can be used in more applications on smartphones.
Examples of image recognition application + 5G
Growing with an image recognition applicationPegamentohas developed a special software where it is possible to take a 3D scan with your smartphone to determine your height. Parents with growth-impaired children can thus take an objective measurement that can be reliably shared with the relevant doctor. A home height measurement has not always proven reliable in the past. The app gives parents and children more control over the exact effect of treatment and saves unnecessary visits to the hospital.
Recognizing artImage recognitioncan also be used to recognize objects. As you may already know from Google image search. However, this option can also be used in real time on smartphones. An example of this is Smartify. Have you been asking yourself for a while what artwork is hanging in the dentist’s waiting room? Scan it with your phone and Smartify tells you who the artist is, what the artwork is called, when it was created and even the materials used. That way waiting at the dentist won’t take so long anymore.
You might also find this interesting to read: the unlimited possibilities with image recognition.
But beyond deploying 5G and image recognition for a driving fish or tracking body height, the possibilities are endless. We discuss these in our white paper. We also discuss various applications in a wide range of industries anno 2021. Curious? Take a look, after all, it is completely non-committal.
Free: Image recognition for dummies
Everything you want to know about image recognition.


