Ultinous Video Analytics Platform (UVAP) is a set of software services that can be composed and extended in a flexible way to build scalable, distributed video processing applications in multiple domains such as retail, security or elderly care.

Video Analysis Capabilities

UVAP provides a rich set of deep-learning-based advanced video analysis models. These models are industry leading in accuracy along with efficient processing. The available core models are the following:

• Head/face detection
• 3D head pose
• Face recognition, re-identification
• Full body re-identification
• Age
• Gender
• Human body pose
• Tracking
• Pass detection, counting

Example Use Cases

UVAP has been used to solve real world problems in different domains. For example:

• Queue management system for retail stores. Customers are counted at the doors and in the queuing area in real-time. A prediction model can tell if there will be a queue in the next few minutes and staff is alerted before it happens.

• Fall detection in elderly care homes. The system uses two 3D-calibrated cameras per room and runs human body pose estimation. Based on the body pose, fallen body pose can be recognized. An alert is sent to the staff in real-time. Staff personals can look at the alert, validate the images and decide to take action.

• Measure waiting time on airports. Face recognition is applied at the entrance and exit points in real-time to measure actual waiting time. Prediction is applied and displayed to customers entering the queue.

• Recognize unsafe escalator usage. Based on head detection, tracking and full body pose, different unsafe situations are recognized. Such as: wrong direction, leaning out, crowd