Google

Google introduces Open Images Dataset: Another innovation in Machine-Learning

Machine Learning in Computer Science and related fields are highly required for the development of next generation Artificial Intelligence based things and machines, including the image identification. Still there are some AI powered websites and services which are able to identify a few elements from an image for machine learning. Google is also now contributing to the development of Neural Networks and Artificial Intelligence based Machine Learning. A few moments ago , Google had introduced their Open Image Dataset which is licensed under Creative Commons Attribution.

Example for Open Image Dataset Annotation from Google

Example for Open Image Dataset Annotation from Google

The new tool, Open Image Dataset is open and free to use under CC license, and it can auto-caption images quickly and clearly without making many mistakes. I uses many neural networks to identify the elements from the picture. For identifying and learning the image objects, Google uses approximately 9 Million URLs to the images that have been set annotations to learn and identify images. According to Google :

the labels cover more real-life entities than the 1000 ImageNet classes

 

Google’s Cloud Vision API too uses a string of neural networks and AI brains to identify the appropriate images and OCR (Optical Character Recognition – aka Image to Text) from pictures which are being uploaded to it. Google makes use of many Human validators to monitor and verify the annotation and remove/add suitable annotations on each image. Google’s DeepDream – A network for visualising neural networks, can also make use of Open Image Dataset. Google is expected to implement the Open Image Dataset on their most of the products including recently announced Youtube-8M and other services like Google Image Search. As the Google have licensed the innovative movement under Creative Commons, it would be available publicly for Developers after some tests.

Source

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

To Top