One of the technologies we’re using at In The Pocket to make digital products smarter is computer vision.
One of the funniest scenes in the acclaimed HBO serie Silicon Valley is when developer Jian-Yang creates an app called ‘Hotdog or not’. By showing the camera on your phone food, it can decide if it’s a hotdog or not. It’s a silly example of one of the most powerful features of machine learning: computer vision. It’s the science that enables software to recognize and understand images and scenes.
One of the most important aspects of computer vision is object detection. It’s the capability of software to locate objects in an image and identify each object. By ‘feeding’ your machine learning algorithm images of an object or person, you can train it to recognize something or someone. A good example are the bike detectors which count how many bikers are passing in a specific street.
Technology can not only recognize objects, it can also classify them afterwards. By feeding an algorithms lots of pictures, it can be trained to recognize specific objects or persons. A perfect example of this technique is the work we did for Telenet: we created an application that can help people with the installation of their Digicorder because it knows which cable you show it.
Quality assurance & security
A lot of industries are still dependent of human judgement when it comes to quality control. Unfortunately humans make mistakes - they can’t focus all the time or often have different quality requirements. That’s where quality assurance with machine learning comes in! Automating quality assurance by using an autonomous, self-improving AI solution can drastically save time and resources.
More and more, your smartphone camera functions as the keyboard. Instead of giving commands by typing them, you can perform a specific movement in front of the camera, which triggers an action. This feature has a huge potential in the manufacturing industry or other places where people aren’t always able to use their hands to control machines.