How will technology be used to manage FIFA World Cup crowds in Qatar?

Crowd control – one of the biggest challenges Qatar faces at the tournament

Researches are trying to find the best solution to this issue. It seems that Qatar’s research teams have found a new technology to help crowd control during the much-expected FIFA World Cup happening later this year.

Considering impaired visibility and density of the audience, crowd control inside and outside the stadiums is essential for ensuring the safety and efficiency of the event.

The answer lies in cutting-edge technologies

The team from the College of Engineering at Qatar University (QU) proposed the usage of cutting-edge technologies which refer to techniques that employ the most current and high-level IT developments.

Cutting-edge technologies include surveillance drones, ICT (Information and Communications Technology), and AI to manage an expected 1.5 million visitors.

The primary focus is set on maintaining the security and safety of participants, fans, and other parties taking place in the FIFA World Cup, said Qatar’s organizing committee. Having that in mind, the state-owned university has come up with an intelligent crowd management and control system in partnership with the Supreme Committee for Delivery and Legacy (SC).

The system consists of numerous components such as crowd counting, face recognition, and impressive event detection.

How does it work?

Drones are providing data for crowd counting and making scaled neural networks to extract significant features and estimate crowd densities.

A brand-new dataset – Football Supporters Crowd Dataset (FSC-Set) – will be introduced during the tournament. It covers 6000 manually categorized photographs of various settings with tens of thousands of people gathering in or near the stadiums.

The research team has also developed a face identification system based on a multitasking convolutional neural network to register faces in various poses. It combines a posture estimation approach and a face identification module. They specifically used the cascade structure.

CNN-based posture estimation method uses the left, right and frontal side captures of faces as the training data. Three CNN architectures -VGG-16+PReLU left, VGG-16+PReLU front, and VGG-16+PReLU right -will identify faces based on the estimated pose too.

Two methods are presented in order to avoid and eliminate unnecessary face information. One is a skin-based face segmentation approach based on structure-texture decomposition, and the other is a color-invariant description. The suggested approaches beat similar state-of-the-art systems in empirical evaluations on four existing face recognition datasets.

Thanks to drone-based video monitoring, AED (abnormal event detection) has recently gained popularity. Drones equipped with cameras successfully detect aggressive behavior in crowds.
Considering all the informations above, it is clear that technology can be highly useful during the World Cup to keep an eye on the crowds outside stadiums and other public places.

Striving to achieve this efficiency, the research team under the direction of Prof. Al Maadeed created a revolutionary AED that seeks to learn abnormal movements and distinguish them from normal.

Abnormal events are learned utilizing a deep multiple instance rating system that makes use of training video sequences with poor annotation. That way, it’s possible to avoid annotating odd movements in training video sequences. They are added to entire movies rather than individual segments.