Now under development: Motion detection, Parallel Object Tracking

Object recognition, motion detection

The cheap and efficient object tracking and object recognition is one of the most important expectations of our age. There exist several methods for the object detection, object tracking and identification. Many of these methods use relatively expensive technologies and easily can make mistakes too. What is certain is that the computer vision algorithms works in a very different way than the human brain, and our knowledge from the latter can be superficial.

In economic terms, it seems obvious that the object analysis can be carried out from the images and from their streams.
The similarity between the computer vision and the human vision is the requirement of light for the visibility of the contrasts.
Sometimes we need to pass a frame through dozens of filters to get that kind of image level which is unreadable for the human eyes. We can demonstrate the properties of the objects through the properties of this kind of images, which is based on the complexity of the decision trees. With this technology we can achieve the following procedures, systems:
- Plate recognition system.
- Motion detector and alarm system.
- Traffic Counting.
- Face recognition based control system.
- Speed measurement, area calculation, volume calculation, direction control, color recognition, logo and plate recognition
- Finding and detecting any kind of object which is able to be described with typical characteristics
Using all of these systems above, we can procreate for example the full automatic vehicle.

As can be seen, the potential uses of the object recognition are practically unlimited. So for given problem, probably we cannot give a boxed, ready to use program, but we can make the program integrating our modules compared to the problem. With this procedure we can resolve the problem quickly and cost-effectively.

Below are some videos about results and tests of our projects:
A few moments from the development.
Demonstration videos.