In this final post of my Azure Cognitive Services week of Azure Every Day, I’ll focus on the Vision APIs and Services within this stack. Azure Vision Cognitive Services leverages image-processing algorithms to smartly identify, caption and moderate your pictures. There are 3 APIs and 3 Services available in the Vision stack:
1. Computer Vision API – Allows you to distill actionable information from images. You can breakdown a video or image and get insights into the content. This is being used in quality control scenarios in areas like manufacturing. In a process where people are putting eyes on an object to say if it’s good or bad, by using this in a scoring method in machine learning, that process can be automated.
2. Face API – With this we have the ability to detect, identify, analyze, organize and tag faces in photos. This can come into play in security. Or Uber uses it in their business app to recognize their drives when they log in.
3. Emotion API – This is pretty cool. You can personalize user experiences with emotion recognition. This API allows us to identify mood or emotion; there are currently 9-10 emotions handled within this API. This may be great to add on to a customer service app or maybe to use with your kids!
4. Content Moderator Service – This uses machine learning to automate image, text and video moderation, so it can detect and remove offense language in a video or in text that identifies an image or even offensive actions in a video.
5. Video Indexer Service – This service allows us to turn video into analytics. It breaks a video down, so you can identify people, work with speech sentiment based on how people are talking, and it will identify key words. What’s cool is it identifies where those occur within the stream of the video. Therefore, you can look back at something you’d like to see talked about in a video and Video Indexer will tell you in what places that occurred.
6 .Custom Visual Service – Easily customize your own state-of-the-art computer vision models for your unique use case.
