Crowd management using AI and ML

Bedrijfsnaam: 
Solita
Bedrijfsomschrijving: 
Solita is a community of highly and widely skilled experts geared for impact and customer value. All we care about is creating impact that lasts, and we have the tech, the data and the insight to do just that. Our community of experts work together with you to help design and build not just any future, but the future we all need. Our unique service portfolio seamlessly combines expertise from strategic consulting to service design, software development, AI & analytics, cloud and integration services. We are a fast growing community of almost 1,000 experts in Finland, Sweden, Denmark, Estonia, Belgium and Germany.
Straat + nummer: 
Philipssite 5B
Postcode: 
3000
Stad/gemeente: 
Leuven
E-mail: 
jeroen.wanten@solita.be
Stage
Inhoud aanbod voor stage of werkplekleren: 

Monitoring a crowd during a large event is never easy. Operators need to constantly  monitor different camera feeds, Will the operator is focusing on one feed, it’s easy to miss an incident somewhere else.

We want to help out by combining AI and ML to monitor in near real-time all feeds to determine crowd movement, density and predict explosive situations. It allows to guide the correct people/teams so the crowd can enjoy the event without any major distractions.

By using audio and visual indicators the operator will be notified of a situation.

Challenges
The main challenges in the crowd counting task are the many variations of appearance, perspective, illumination, crowd density and distribution. In terms of performance (speed and accuracy) traditional AI solutions for detecting objects in images suffer a lot when it comes to the detection of too many objects per image.

Applied Methods
To approach the problem of varying crowd densities and distributions and the high number of objects per image, a density-based approach is considered. In effect, the prediction target of the AI is the density of people on a per-pixel basis. For this purpose a neural network architecture, called CSRNet, was implemented. The network is composed of two components. A convolutional neural network (CNN) is used for extracting features in the images. The second component is a dilated CNN, which uses dilated kernels to deliver larger reception fields.

Gewenst profiel: 

1.1Your Role

  • Research different AI and ML platforms, keeping scalability and HL in mind
  • Get to know the basic concepts of crowd management
  • Learn about predictive analytics, AI and ML technics
  • Integrate different platforms to create value-added solutions
  • Become an (API) expert 

1.2Responsibilities

  • Improve your own and Solita’s knowledge
  • Explore the AI and ML landscape
  • Design your own solution architecture with the help of your colleagues
  • Present your solution to our Belgium colleagues

1.3Profile

  • Result oriented and creative 
  • Good analytical and problem solving skills 
  • Eager to learn, grow and take leadership 
Opleidingsprofiel: 
Bachelor in de elektronica-ICT: ICT
Bachelor in de toegepaste Informatica
Graduaat in het programmeren
Hoe kandidaat stellen?: 

Send your resume and motivation letter to jeroen.wanten@solita.be

Gewenste periode voor stage/werkplekleren: 
all year long
Deadline: 
vrijdag, 9 oktober, 2020 - 11:30 tot vrijdag, 9 april, 2021 - 11:30