"Angus.ai" which can eye track the line of sight of the person who came to vending machine and analyze age, gender, mood

byinefekt 69

By attaching a commercially available web camera to a vending machine, it is possible to track the age group, gender, mood, eyes movements etc of the purchaser in real time and analyze it as data when a drink or food is purchased. AI "Angus.ai"Has appeared.

https://www.angus.ai/

You can see what Angus.ai does with the following movie.

Angus.ai - Audience Analytics - YouTube


A man who installs a commercially available web camera in a vending machine.


Activate the camera, activate the application.


Preparation is OK only with this. The situation around the vending machine is captured 24 hours a day, 7 days.


The algorithm of the application grasps the number of people who passed the vending machine, the number of people who stopped, the downtime of the person, and the movements of the eyes.


In addition, users are recorded with information on three categories: age category, sex, mood.


KPIIt will be visualized on the dashboard in real time.


The dashboard looks something like this.


Also, data is collected through HTTP REST API.


Three pieces of information about categories, gender, and mood are written for users, but you can see how the algorithm understands the mood from the following movie.

Angus.ai - Emotion / Age / Gender Detection - YouTube


There are two men and women in front of the camera. If the face is surrounded by a frame, "39-43" "ID: b9ba" "emoticons emoticon" is written on the blue label if it is male. It seems that the color of the label indicates sex, the number in the first row indicates age category, and the emoticon is mood.


When the two of us opened their eyes, the emoticons changed to a surprised expression.


The woman kept a surprised expression, only a man smiled. Emoticons also keep track of changes in facial expression properly.


When I got angry expression, I showed the mood of anger with emoticons.


Furthermore, with Angus.ai's algorithm, it is also possible to grasp lack of goods shelves in real time. It will be possible to operate the store semi-automatically by this.

Angus.ai - Real time Out of Stock Detection - YouTube


When someone reaches for the product on the shelf, monitoring starts with "This product may be out of stock". The part surrounded by the blue frame is being monitored as possibility of missing parts.


If someone reaches out to the last product and a missing item comes out ......


Immediately missing items are detected and the blue frame changes to orange.


When the store person starts replenishing the product ......


The orange part returned to blue again like this.


In addition, we track not only the product but also the movement of customers.

Angus.ai - Smart Notifications - YouTube


What is useful for tracking customer movement? ...


The algorithm tells us that you are about to wander the same place or stop at one place, "I need help."


Of course, it will detect missing items as well as the previous movie.