Google have just announced that they have open sourced their object and scene detection API. This is a game changer as it will supply the masses with the best quality solution on the market and all completely free. If this is the case, then why are some companies still paying crazy amounts of money for this solution?
We all know that “visual” is becoming increasingly important. We are constantly being bombarded with images, gifs, and videos on various social media platforms everyday. In fact, the number of images being shared daily on social media has grown from 300 million images a day in 2012 to over 3 billion images a day in 2016. So object and scene detection are great ways to enrich the visual data that is generated.
Over the last ten years Google has built the best in class object and scene detection. But why is Google’s the best object and scene detection technology out there? Well, Google has access to the largest volume of data and, as it is based on deep learning, it is a data volume game. This means that the more data it gathers, the better the object and scene detection technology gets.
According to TensorFlow and Google, this open source framework makes it easier to construct, train, and deploy object detection models. It is also Google’s first open source release and includes models that were trained on the Microsoft COCO dataset. What’s exciting is that it allows users to integrate the API without having to train models themselves, but it also provides users with scripts which allow them to train their own models if they wish. So anyone can do this on their own without an expensive subscription to an object and scene detection provider.
By open sourcing their object and scene detection API, anyone can integrate Google’s models directly into their product or products. To find out more about how to integrate Google’s object and scene detection, or to request a free Gold Standard Logograb logo recognition API test, CLICK HERE.