Google is all set to release TensorFlow Object Detection API that will help the developers in real time. This new set of intelligent detection capabilities will assist the researchers and developers to discern the object in any image. This will be the part of Google’s ongoing development of TensorFlow Framework and will be available in the company’s open source community.
Google is making best of its efforts to boost its performance yet keeping it simple for the users and developers. The released model proved to be quite a robust in the evaluation and can be a regular need for researchers and developers out there.
What Does TensorFlow Object Detection Do?
TensorFlow Object detection API hands over scientists and researchers the access to the exactly similar technology that the company uses for its own system. For instance, the tech behind the Nest Cam, the same set of image search and that street number identification in Street View. Moreover, the system that Google launched made a win in Microsoft’s Common Objects in Context (COCO) object detection challenge the previous year.
The set of models released today in detection API includes inception-based convolutional neural networks and streamlined models to work on not many sophisticated machines. Also, MobileNets detector comes in handy, optimizing to run in the smartphone for real worldly object detection.
Previously, the company unveiled the set MobileNets detectors that have the ability of facial recognition, landmark recognition and object detection. The company introduced these models as lightweight computer models.
Now the smartphones do not tend to have server-based setups and resources of larger scale desktop. This leaves two options for a developer, that is machine learning models or simplifying the models themselves. Though machine learning model is suitable for cloud, it appends latency and needs an internet connection. Whereas, the other approach is to make a trade-off for the sake of more pervasive integration.
Other Tech titans Are Also In the Same Business:
On the other hand, Apple and Facebook are also making efforts to contribute resources to the similar mobile models. Last year, Facebook launched its Caffe2Go framework. This program was to build models that can run on smartphones. Also, this was the first mega implementation of Facebook Style Transfer. Moreover, In Google’s I/O, the company introduced its TensorFlow Program. And in the recent time, Apple in its WWDC unveiled Apple CoreML. Apple’s attempt for the developers to lessen the technicality of running machine learning on iOS.
It should come as no surprise that Google hits the shot that reaches farther than attempts of Facebook and Apple. It is because Google public cloud offerings that keep its TensorFlow program ahead of others.
Here is the model of TensorFlow Object Detection API released today.