Event venues challenges, how to improve operational efficiency?

What are event venues challenges and how to gain improved operational efficiency


Sports arenas, stadiums, concert halls, theaters, conference venues.

Challenge: at-home viewing experience of mass events like matches or concerts is in many ways better than live presence at the event venues:

  • Affordable: first-row view is offered at no or low cost
  • Time-saving: no traveling to the venue, no traffic and parking issues, no waiting in long queues to and out of the venue and to facilities inside for snacks and drinks
  • Comfortable: pause upon request, restroom available on demand, no disturbance from undesired behavior of other participants
  • Healthy: no exposure to virus and bacteria in the crowds, adjustable sound level, no external weather conditions
  • Secure: no risk of unexpected aggressive behaviour or panic
  • Accessible: suitable for impaired and elder people.

The need for live experience is not dead, though. It still can beat the comfy living room for its uniqueness, all-sense stimulation and emotions.

But in order to successfully compete for the fan, event venues need to adapt to growing demands of the audience, by trying to improve their operations in the critical aspects.  As labor cost remains a significant factor in the operations, the attention turns towards technology and automation.

So, how technology may help?

Facilitate on-boarding to the venue:
  • Computer vision supported parking control systems: monitoring the occupancy and assisting the driver to easily reach the free spot.
  • Computer vision supported queue management: live monitoring of queue lengths, redirecting fans to optimize the flow.
Control the inside traffic:
  • Computer vision: redirecting fans to optimize the flow, providing predictions on waiting time in the areas of interest.
  • Detect fans that disregard sanitary regulations (eg. face mask detection), prevent them from entering the venue.
Security surveillance:
  • Computer vision: monitoring the crowd, detecting unusual behavior, detecting prohibited items (guns, knives), facial recognition upon the entry (eg. excluding fans with stadium ban), detecting slippery floor.
Visual experience:
  • Computer vision: ensure visibility of details at all spots, enhance the camera view with visualizations and supplementary information (eg. ball tracking, speed estimation).
event venue concert hall stadium

Involve advanced technology

All of the above technical solutions involve smart computer vision algorithms that are able to detect and classify objects on the live video stream and take actions based on the detection output.

The design and development of such a smart algorithm begins with setting up the scope: what objects need to be detected, in what environment we want to detect them, what are typical conditions and what extreme situations may occur. These questions help to frame up the expected solution and development process and allow for setting up the requirements for the algorithm to be trained. It is recommended to include a Machine Learning Specialist at the early stage of the solution design, as this is the time when detailed decisions are taken on preparation of the training dataset. 

As the AI algorithm is only as good as the dataset it is fed with, decisions such as quantity and type of classes of objects to be detected, and the types of relevant attributes of the objects and the scenes may impact the quality of the final solution. In this stage, the experience and domain knowledge are a valuable asset that may save the costs and reduce time needed to develop a working application.

parking lot event venues

How we can help?

Task: prepare dataset for parking lot occupancy monitoring

Our consultant will assist you to set up the project scope and requirements by discussing these aspects:

Understand and state the goal:
  • What objects do we want to detect (is it necessarily all types of vehicles?)
  • Do we want to track the vehicle to recover its journey, re-identify in different places
Understand and describe the environment:
  • open street? underground car park?
  • lighting conditions,
  • image sources (camera types, views, distances).
Discuss and define dataset requirements
  • variability in conditions, views, objects,
  • quantity of images,
  • annotation „density” (single car per view, multiple cars?)
Define classes to be annotated:
  • single class vs multiclass,
  • attributes per each class, attribute values.
Document and sign up the requirements
bobox consulting project management

Here come the PM and annotation experts:

Take first iteration of annotation

  • Collect issues
  • Discuss decisions
  • Record Edge cases in documentation
  • Hand over first batch for verification

Continue with next iterations observing model improvements