Last week, Google released first beta for Google Vision API. This new service extracts classification information from images by using several methods:
- Label detection (from Google catalog)
- Logo detection
- Landmark detection
By now, this API is only available in English, but there is no problem if combining results with Google Translate API.
You have to note that this services are not completely free, a valid registration via credit card is required at Google Cloud Platform. However, invoicing starts at high volumes, so developing and testing tasks can be performed without paying.
In keensoft we’re starting to test automatic Google Vision integration with Alfresco by using rules. At our Github is available our first addon on this: alfresco-google-vision, which provides labels, logos, landmarks and OCR for any language included in Google Translate API.
Our first experiences has been very interesting, as it can be shown.
We’ll follow Google Vision evolution in the next months, because this product is going to be required in the future for every image auto-classification task on ECM.
April 28 & 29 happens BeeCon 2016 in Brussels. The second day, Axel Faust will lead a Hackathon and we are proposing different evolutions for this addon to be developed during this meeting. Come on and join us this Spring in Brussels!