Have you seen machine learning on the cloud platform?

 

How exactly it works right You're going to see that part in the programmatically That's what you're going to do Just for recognition We'll use an A p a call on I'll show you how exactly you can make an A p a call but remaining services we're going to see through only green mode because concerning the time what we have fair enough does that answer the into the Christian Yes in the images can mean any former right now it's opposed JPEG PNG waters format Do we have not many formats because that's what I'm not sure about the last two months was there any change Because the reason behind me telling is AWS environment Even if you're an expert what's really going to happen is every single week they come up with a new release So they try to improve based on customer um comments.

You never know whether they started supporting the remaining parts That so that's a reason for it But when I left it was just too for my supporting Okay so we're just going to close this on I'm going to go back first let's see through the slide one's See the recognition part, Yeah So extract rich moderator from official government objects Seen facial analysis face imbalance and facial recognition and tell ability Recognition So let's play around with agree on the part first and then move on to the programmatically Okay Is there a quality check for the image like not blur or something like that.

No there is now much of a quality check but that data is what you're going to get from here You can see how much confidence it is based on this data you can see Yes It's how perfect That image is actually may have toe God make this method call on Figure out Is that a shaky image or was he able to do a prediction All those methods and functions are available over here and this is nothing but a response Um that's been sent back given by face call face compared app or something like that which I'm going to show you in the short Fine Okay the next Which run the image What about you Sorry to interrupt him in one question Say for example that is that's an area where no direct twins No more or less.

The two identical people Yes If that is identified I mean then what kind of person did you tutor has done not doubt that would react Oh so if that print is identical twins it's the same matching It's going to pick up and say both these matches exactly like a 95% of your 92% age Yes it does the same matching for both this in majors OK it's going ok Okay Thanks Yeah Okay So the next five days object and scene detection Right So you have seen Google car and all those the car detection How exactly has been then So it needs to figure out in an image Whatever the majors are given is that a car is a human is escape board on This is a confidence interval You can see over here the confidence interval with which it's doing this prediction Let me just show you that Paris felt.

Let's upload one of our images on dumb Yeah its, so the next question from someone was health were different Is it from Microsoft Please FBI So every single floor Pereira was trying to eat up in this particular area Right But this is machine learning on the cloud platform is going to be the next trillion-dollar business So every single player is trying to make they're correct to make the service as accurate as possible And that's there That's the that is doesn't happen right now right So Google is trying with their own innovation Amazon is on track Honda even um Microsoft in on their own base So, so one Al Gore them being made better The next one is trying to improvise on it So identified customer is gonna definitely get benefited with that So that's all I would be able to see It's almost saying but I have not played around much of the Microsoft face A p A bit I hear from customers.

It's almost the similar kind of functionality that we have So some of the features made before here and there but at a high level that's going to be the same Okay let's pick up this wreck Hyphen pick Um that's the A major issue are just predict Um that's pretty straightforward It's enamel ma'am I mean animal mammal because I'm not sure what exactly when you called horse and styling and all those things Anything about 95% age of 90% 90% of prediction That's exactly what he wanted to pick up for your application-level So if you're going to write the senior application you're going to say Yeah I'm pretty much confident that this is a horse ongoing to use an application.

Some of the application where you can see this is you don't need to have He can create your own collection I d on Ah for your organization Basically you don't need to swipe your badge to adore instead What you can do is recognize your face and automatically is going to open So So I was there with an accent you almost five years back So we had a similar kind of project that isn't running So it took almost three months with around 40 f d working on it right So that our requirement was um When does the face has been recognized It's going to say hello Shari on then Then the door is open for you And that's the kind of application that we want to build And now using this sort of technology it may take you hardly 2 to 3 days to create exact similar kind of application right.

You don't need to swipe your card So that application was basically engaged An intruder is found in your from ISIS It's gonna immediately back that particular face and it's going to show it in the security managers mobile phone It is going to say I see an intruder here What do you want to accept or reject this particular in Judah Right So the moment that security guys going to say reject it's going to buzz an alarm and then you may have to Some security is going to come and escort you and then take up another I B and things like that So that's how those applications can be built It I also some of the good use cases for this would be how to pamper those customers very, very privileged When you walk into that's the kind of application that is being billed for some of the ah A domains right now and pick up Amitabh Bachchan or anyone.

It's going t recognize all those images famous personalities Everything is going to get detector Deacon to use facial next animate This might be another use case where you might be ready much looking for So anything in this particular made has been picked up and totally here You can see over here This might be a very good use case into this right The total boot whenever the card number is being flashed immediately can pick it up and you can see over here See j 389 and under the oldest things, I've been captured over here right.

These are the kind of use cases Oh where you can use Um Recognition service Okay let me just real quick show you how to use this and programmatically as well Let me just open up Atom Okay Detect face Detect Okay So what you're trying to do The one thing what I have done through the gooey option So us an automation guy is a programmer You may not be using our money as our detective Even I would be interested in logging every time And then he wanted to have an automated system in place where he's just going to feed it Okay The moment he gave him the picture.