APPLY Case Studies in A.I.


We have used computer vision and predictive & behavioral analysis to develop A.I. solutions for 7 industries. Here are some of our case studies.
Media

AdEx - Real-time visual media analytics. 

A.I. and Neural networks automatically analyse content from video streams (cable, satellite, IPTV networks or OTT platforms).

AdEx is able to:

  • automatically detect ads from TV stream in real-time

  • identify and detect various objects, subjects and scenes

  • extract text from screen and speech-to-text

AdEx is currently used to track & analyze TV ads. It means that media buyers can directly track if their ad have been aired at the right time and in the right amount. AdEx also allows to compare your competitor ads side by side with yours. The system also uses big data from broadcasters so the clients can see live viewer statistics for the time of their ads or content. This technology also is used to track politician activities in election period as well as monitor propaganda.

In short - finally there is transparency!

Industrial

Industrial grade robotics adopted for nonlinear tasks.

One of the biggest challenges for faster and wider integration of robotics into everyday manufacturing and servicing tasks, is the entropic and unpredictable environment of the real world.

For us it is easy to figure out which object is to be picked first from a box of randomly arranged objects of different shape and size. For traditional robotics - not so much.

We are talking about an A.I. vision enabled robot, that can autonomously analyze contents of a container by determining its shapes and positional relations of different objects. The system also is capable of classifying objects based on size, weight and visual appearance. For each type of object a robot encounters, there can be a different routine.

This product is intended to make significant steps towards changing the status quo in the field of robotics.

 

Smarter robots for metalworking.

When working with large-scale raw metal, it is normal to have variations in dimensions. When a human worker is welding or cutting such a piece, there is no problem at all. However, additional programming and checks by humans need to be performed if robotic manipulators work on such metal. This can be time consuming and since there is human evaluation involved – prone to mistakes.

In collaboration with ABB Robotics, we are adapting standard robotic manipulators for automatic evaluation of raw metal and synchronizing it with the blueprint to achieve identical results every time.

For example, the system scans a metal beam and detects that it is wider in middle. It then modifies the plasma cutting profile to take that into account.

Finance & Insurance

Automatic medical receipt submission for insurance company.

We developed a solution for one of the largest insurance companies in Europe “If P&C Insurance” LTD.  The solution combines computer vision and optical character recognition.

The problem.

So far, clients have collected their medical receipts for a lengthy period of time before submitting them. The main reason for developing a new and innovative solution was the submission process - it took a lot of clients’ precious time.

The solution.

The client takes a photo of the receipt, uploads it to the insurer's system and that is it.

Our A.I. system:

  • recognizes a receipt in the uploaded image

  • understands the information is in the receipt

  • fills out the form automatically

The best part is that the A.I. can extract information from all kinds of receipts.

Casino

Vision system for casino table accounting control.

This product is a part of Casino Management System (CMS). It performs visual observations of any card table and tracks chip exchanges between dealer and players.

As part of CMS, it tracks whether dealer makes correct payouts to players by analyzing both -  card combinations and chip value.

Also, it builds statistics for each dealer - their performance under different circumstances.

 

Vision system for fortune wheel type game monitoring.

System for monitoring of fortune wheel type game. It visually observes the wheel and determines the start of the game, wheel speed, direction, game stop and winning sectors. It directly uploads all the data to the casino server for processing and analysis.

The system can work with any type of video feed, so it does not necessary need to be located near the wheel itself.

Microbiology

Artificial Intelligence enabled dairy product contamination control system.

The current industry standard for dairy product quality control was developed in the 19th century. It is simple and fool proof - take a piece of a dairy product, put it in a petri dish with food for bacteria and allow fungus to grow. It takes on average five days.

Drawback - you need to store large amount of processed and consumption-ready food before it can hit the shelves, thus shortening shelf life significantly, spending a lot of money on refrigeration. And we are yet to add the logistical costs.

Our product will change the 5 days waiting period to a couple of hours. It is achieved by using a patented contaminant cell biomarking method and an automated microscopy platform.

This platform scans the whole sample, finds actual cells of microorganisms and provides exact counts of them. Visual recognition is done by utilizing neural networks. In fact, this system can be adapted to a wide variety of microscopy visual analysis cases.

Healthcare

Interactive face motion game to support children with speech and face disabilities.

CheeksUp - 3D facial expression recognition system, that gives feedback to users in real time through game-like animations.

The idea is to engage children in otherwise boring and repetitive tasks of facial muscle exercising. All exercises are adopted from traditional speech and facial muscle rehabilitation practices.

This cloud enabled system also includes patient & therapist profile system, thus it tracks detailed activity and progress of patients, generating useful medical history data.

Surveillance

Detecting fire before it starts, or spreads.

A company that owns multiple office buildings needed a solution that can detect potential fire breakouts at a very early stage. We used computer vision and here's why. A camera can see things the human eye cannot - infrared and ultraviolet light.

The A.I. can detect the difference between a flame from a candle, or a fireplace and a smoldering curtain or carpet.  It does it much faster than standard smoke or flame detectors.

Combining the A.I. with advanced smoke detectors, determines the difference between cigarette smoke or a curtain on fire.

The result - a sophisticated system for hazard alerts.