Azure AI Mentalist

this is the panel for user to see the change of people emotion
this is the architecture of our project

Recently, there is a huge growth of news all around student commit suiciding cause by mental issues or pressure. Those news breaks everyone hard. So, we think, how can we spot the needy student earlier and help them. That’s why we have Azure AI Mentalist, a solution which can benefit from education, special education to commercial activity.

At tradition high school, one teacher needs to face around 30 students per class. But at university classes, a professor needs to face more than 200 students per class. So, it this so hard for our teacher to find any potential mental illness student and help them.

Our solution allows the teacher to use any video recording tools, such as a smartphone. Shot a class in multiple angle views. Put it on to cloud use AI deep learning analyze facial details and emotions. We can get the result as soon as 2 minutes. The teacher can see the statistic after class via the web panel, to see how student’s mood change throughout the lesion, and spot any student change of mood. So that we can help the teacher to improve their teaching quality. Other than that, our system will gather emotion statistic for every student throughout the year as a big data, the school can easily spot any potential mood disorder student, to give help as soon as possible.

On the other hand, Hong Kong has been promoting Integrated Education for a long time, but normal school teacher’s lake of skill handling special need students. Our system can help special education worker made-to-measure courses base on student need. Such as Autism and Attention Deficit etc.

Last, we jump out of the box, using the relationship between emotion and event, we can fit our system into many commercial cases. Such as hotel reception, we can place camera facing both client and receptionist. To reflect services quality. And another example is outdoor commercial, we can see if the commercial is success base on walker’s reaction. The traditional commercial system uses feedback letter to collect user reflection, but the real answer is deep in our mind. Which we can’t simply write on paper.

In our architecture, we need some compromise and breakthrough, Azure has the new video indexer, but it lacks emotion detection. But with Emotion and Face API, in a nutshell, it won't support videos. Our system cleverly uses their own benefit and work together, make videos has emotion, but still outperform video indexer, which is shockingly fast.

What we’ve learned is to take advantages of the high connectivity of cloud services, to make an impossible solution possible. And this is Azure AI Mentalist.