Faces

Sign in to queue

Description

In this episode of 'Context', we take a look at facial technologies and how we can make use of face detection and recognition in modern, Universal Windows Apps and also how we can use some of the additional Microsoft Cognitive Services for other functionality and cross-platform work.

Here's a breakdown of the show;

Show Notes

Please note that the recording refers to 'Project Oxford' - these services were moved to the 'Microsoft Cognitive Services' shortly after the episode was recorded but at the time of publication this renaming has not reached through to SDKs and so all the hyperlinks and code samples should work fine.

Embed

Download

Download this episode

The Discussion

  • User profile image
    Ian2

    I think you should have titled this one 'Mugs'. 

    Must admit quite impressed by Kinect which seems to pick me up when I am just strolling past.  Looking forward to Cortana integration for XBOX  - facial recognition in family (or group environments of known people) might throw up some new/interesting opportunities.

    Project Oxford looks pretty slick, damn, there are not enough hours in the day ...

  • User profile image
    Polotto

    Amazing video!!! Can you make a video testing the face recognition with pictures? I have a doubt if the face recognition knows the difference between pictures of people and real people.

  • User profile image
    mtaulty

    @Polotto: thanks for the feedback. In terms of the question - the depth cameras should not be fooled by a photo as they do not work that way, it's a 3D view that they are capturing.

  • User profile image
    hal9k2

    I laugh when you said "App will crash" when Andrew was taking picture in one of the first demos :D

    BTW if you taking picture without depth information (like in project Oxford demo now 'Microsoft Cognitive Services) can you show printed picture and be recognized ?

    Great show, great topics. Thank you.

  • User profile image
    mtaulty

    @hal9k2:Hi, what's being sent to Cognitive Services over the REST API is just a 'flat' photo and so the service is able to tell you who it thinks it sees in the photo but it can't know much about how that photo was taken. So...I suspect if you submitted a (quality) photo of a photo then the service would still perform recognition on it.

    BTW - glad we made you laugh :)

Add Your 2 Cents