Funny Object Recognition Fails
I am in the process of creating a video review of this amazing objects detection app "Aipoly Vision". As I start using it, like with many object recognition apps, there are some funny fails. So in this video, until my full (serious) review is ready, I wanted to share some of those bloopers with you, those that happen when I tried this app this morning for the very first time.
Keep in mind that this app is much smarter than it is shown in this video, and it (the image detection algorithm) continues to be fed with more objects and images so it can recognize more objects and do that much more accurately over time. You can unlock more categories that help the app recognized more objects (I just used the general icon). By choosing a category, you allow the app to make more accurate "guesses" and many of the categories are unlocked, and you need to subscribe to the app in order to unlock them.
According to the developer, the app can recognize thousands of things in real time, identified hundreds of colors and learns new items over time.
This image recognition algorithms are built on top of learning algorithms that improve the accuracy of the object classification by showing it thousands of examples of objects and backgrounds, and although there are different learning algorithms, the core principle is more or less the results should allow the algorithm to make more accurate classifications, in other words, describe the object more accurately the more it "learns".
So enjoy these bloopers, and a full, more serious review will be released in the near future. The app has many features that I want to talk about in-depth so it will take me some time to make it. Even with the very short time I used it, it's one of the more impressive and useful object recognition apps that I've tried to date, so I can't wait to use it more and share my full review with you - stay tuned.
Emoticon icon attribution: Emoji One [CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0)]
#ObjectRecognition #ImageRecognition #ObjectDetection