Hello guys. Welcome to the 42nd session of using Microsoft Word framework Luis and cognitive services tutorials for beginners. In this session we'll learn face APA. For this session, I will use APA from project expert because in Microsoft Cognitive Services APA basically used to recover is limited to. Using this we can detect faces within the image. We have already used the Vision API to retrieve faces.
But these face API provides more details about face. We can identify face attributes such as age, gender, hair color, person has glasses, beer for stretch, and help me cook etc. Even we can often face clamor, which includes coordinates of the face within the image Along with these, we can also find coordination for lips, nose, eye mouth, cetera. Face API will generate UID for each face. Using those, we can easily store and compare with existing faces. We can also find similar faces for given face and alternate the group of face easily using face API.
I have already added project Oxford dot face package and added in. In the demo, I have created two more constants for pasting an endpoint that is create new face APA in Azure. Search please click on Create require details I have already created so I'm opening it click on the overview copy and pasting address and also copy key class copy signature of this method and Jeannine this method let's create an instance of face service client recognition API. In face API we also need to pass request features. Here I need age. facial hair.
Gender glasses. Makeup hi Use detect a sync method to retrieve all faces and fast image URL and required features. From this, I'm obtaining attributes of the first face. Hold on Now we already see it in vision a play button face API. We got some more details. They can return details of makeup such as eye makeup, and makeup.
Using facial hair, we can check the face contents. beard, mustache and sideburns. And even we can also check for glasses type that plays on the face that is treated as a string and update dialogue publishing Create a new conversation with the image of Emma Watson. Here we got her age is 23.5 years. She is originally female. And she makeup eyes and lips.
She does not have beard and mustache and does not wear any glasses. But I still like this image of Robert Downey is ages 40 and does not makeup I and you're reading glasses by using free landmarks we can often coordinates of fibro, left eye and right eye mouth nose to print early. In this session, I'm not showing it but if you required then in detects the same pattern they need to pass through for return face landmarks increase APA, we can compare two PCs and verify that both images have the same person or not. For these letters, copy my third signature. Change my attorney that is create an instance of face service clients require two phases and recharging phase one from this URL. Let us see this image contains Robert Downey faced with classes and nominal views And we will detect another face from image that was added by users.
To compare two faces we need to use verify as it takes two g IDs that represents unique face ID. So let's face it from those faces. Using is identical, we can check both faces are same or not. And we can also check for confidence for this comparison that is a bit of a dialogue to seem to find and publish it. Start a new conversation. Here I'm going to use this image.
You can see that there is a various difference between those images, including hairstyle beard, glasses and emotion. We got both faces are identical. Now let us compare it with Leonardo. Got false, which is 0.09 confidence. They just try to compel the face off, I love it. Obviously, both faces are not identical and in one record only 0.01 confidence.
Let us see what we have learned in this session. Using face API, we can detect faceless mass coordinates. As well as we can identify faces gender, and race have makeup theater glasses. We can also use face API to compare two faces. For these we need to pass to face ID. Thanks for watching this video.
If you have any doubts, please feel free to contact me. Have a nice day.