Microsoft Vision API

Microsoft has rolled out a new Vision API which can automatically categorize imagery of an adult nature. The machine-learning API is part of Project Oxford, and like the face detection API announced yesterday — which determines your age and gender based on a photo — is aimed at assisting devs in creating more compelling apps for Windows and Windows Phone.

The Vision API analyzes an image's visual content for "image categories, pornographic detection, dominant color, and more:"

  • Use the adult and racy features to enable automated restriction of sexual content and protect your users.
  • Utilize categorization to help append tags to images, as well as group images into clusters.

In addition to image categorization, the API can generate a 'high quality and storage efficient" thumbnail from an input image:

  • Leverage thumbnail generation to help present images in the best form suited to your needs.
  • Use smart cropping for thumbnails that are different than the aspect ratio of your original image in order to preserve the region of interest.

There's also OCR (Optical Character Recognition), which converts any text found in an image to a machine-readable format:

  • Run on embedded images to generate text and enable searching.
  • Allow users to take photos of text instead of transcribing to save time and effort.

You can try out the features offered by Vision API by heading to the link below and uploading an image:

Vision API image categorization demo

Project Oxford also features a Speech API that offers real-time speech recognition, Speech Intent Recognition — which leverages Microsoft's Language Understanding Intelligent Service technology to understand the intent of the intent of a sentence, speech-to-text conversion as well as a text-to-speech utility that converts written text into audio.

Source: Microsoft (Project Oxford)