Computer Vision

ORGANIZING

One of the critical steps of computer vision is to is to group information by comparing it and to divide it into different classes. We can thus get an idea about the dimension and complexity of the problem we are trying to solve. For example, from a bird view image of a sports game, we can distinguish the referee, the players and the positioning of the ball. According to International Data Group (IDG), by 2022, 93% of the information generated in the digital world will be raw and composed of visuals. For data, the problem is not only the volume but also data not being interpreted. Successful examples for the ‘organizing’ stage of computer vision can be reference catalogs 1, ‘eyes’ for visually impaired people 2, sports statistics 3, security camera calling the police 4 and damage reduction 5.

1 Lens

Developed by Pinterest, Lens is called the ‘Shazam’ of images. Let’s say, you see a bag when walking down the street or a couch in a shop and take a photograph of the object. Then you open Pinterest and it immediately tags the product with certain hashtags and lists similar products.

PROJECT DETAIL

2 MyEye

For a visually impaired person, leading a self-sufficient life is very important. Thanks to a tiny camera attached to glasses, MyEye becomes the eyes of the visually impaired either for reading a book, or finding the ways during transportation or for choosing a product during shopping. When you point your finger, the camera focuses on the spot and the microphone in your ear starts guiding you. MyEye tells you the name of a familiar person approaching you, indicates colors when choosing products and recognize banknotes during money exchange.

3 Sports Statistics

Statistics has become one of the critical contents of modern sport games. Bench staff that used to take notes with pen and paper is history. Now, there is AI taking care of this job. Cameras put on different places in the field collect various movements and visualize them. Indicators such as intensity maps, effort per athlete, success analysis, motion analysis are now sine qua non for a sports game...

4 Camera Calling the Police

With their motion sensors and computer vision, new generation security cameras not only filter out false alarms but also automatically inform police stations if necessary. There are many new algorithms developed for preventing false alarms such as branches swayed with wind, birds flying close to the camera or pets moving inside the house...

PROJECT DETAIL

5 Damage Reduction

In order to completely prevent minor accidents within the city, many car manufacturers have been working on features such as autonomous brake systems. Computer vision applications supported by different technologies constantly observe 360 degrees around the car, categorize details such as people, animals, vehicles, poles and road lines in the environment and estimate their motions.

PROJECT DETAIL