Computers can often be seen as somewhat paradoxical. In many ways the average computer is a super genius. Even the best chess players will struggle against computer based opponents. And there’s no question about the skills computers have in matching data to a text based query. But at the same time there’s been certain areas where computers simply haven’t been able to compete with people. Some tasks which seem simple to people have come very slowly to machines. But every now and then there’s a game changing breakthrough in technology.
The latest example of this can be found within the work of a company called Slyce. They’ve managed to create something that the world has often seen hints of before. But unlike prior implementations, Slyce has delivered the real thing to people. The company is among the first to offer something known as visual search.
Visual search technologies are based around searching for information based on visual cues. This can be best understood by first looking at textual search. This is the kind of search that one would use with a traditional search engine. A person puts a few words or sentences into a search engine and then clicks on a button. It’s a wonderful technology and one well suited to a number of tasks. But at the same time it comes with a number of inherent issues. It’s essentially trying to live by word alone in a world that is heavily based on sight. Imagine trying to identify things with a blindfold on. That’s the big limitation which most search engines operate under.
The big issue is that sight is a far more difficult problem to solve than most people realize. Sight comes easy to humans because people have a huge amount of their brain dedicated to decoding visual input. It feels natural because the process of sorting down an immense amount of visual data to a more usable size is quite literally instinctual. People’s eyes take in visual information, but in many ways it’s the brain that really handles sight.
Slyce recognized this fact and knew that it’d take a large amount of processing power to handle machine based visual search systems. But at the same time, visual search is most useful with heavily portable systems such as smartphones. However, the processors on these devices are hardly up to such an immense task. As such, Slyce decided to try a hybrid approach. The potable system could handle transfer of data to a larger system, which would then send back the results to the small processor. Given the speed of most mobile networks these days it can create a fully seamless process. Slyce has created a system by which apps can actually see the world and offer useful information based on those sights. The technology is also able to be licensed out to work with one’s own apps.