The new SentiVeillance by Neurotechnology provides real-time object detection and classification, biometric face identification of multiple people moving in different directions.
Neurotechnology, a provider of high-precision biometric identification and object recognition technologies, announced the availability of the SentiVeillance 4.0 software development kit (SDK). This latest version has a new name and provides a number of enhancements over the previous version (known as VeriLook Surveillance 3.1) that support its use in a wider range of surveillance, security, and public safety applications.
Working with live video streams from single or multiple high-resolution digital surveillance cameras, the new SentiVeillance provides real-time object detection and classification, biometric face identification of multiple people moving in different directions.
Working with live video streams from single or multiple high-resolution digital surveillance cameras, the new SentiVeillance provides real-time object detection and classification, biometric face identification of multiple people moving in different directions. It includes adjustable motion sensitivity for improved tracking of moving objects and detects if a person or object (such as a car) is crossing a line or entering or leaving a predefined boundary area. This data can be used by software programs to trigger alarms or send alerts to law enforcement or security personnel.
"The ability to create custom object classes, along with improved tracking of moving objects, gives the new SentiVeillance 4.0 SDK more flexibility to be applied in a wide range of surveillance tasks," said Raimond Laptik, Surveillance Research Lead for Neurotechnology.
The addition of camera calibration for ground plane definition enables the SDK to calculate the area and speed of moving objects and allows users to create custom object classes, such as pedestrians, groups of people or vehicles, based on these parameters.
The addition of camera calibration for ground plane definition enables the SDK to calculate the area and speed of moving objects and allows users to create custom object classes, such as pedestrians, groups of people or vehicles, based on these parameters. This enables SentiVeillance 4.0 to detect when people or objects enter zones or cross pre-defined boundaries in the wrong direction, making the SDK suitable for use in transportation safety applications, such as systems that set off alarms when someone gets too close to train tracks or when cars are about to make a wrong turn onto a one-way street.
The biometric identification capabilities in SentiVeillance 4.0 are based on the latest facial recognition engine from Neurotechnology’s MegaMatcher 6.0 product line. SentiVeillance can match face images against internal databases, such as authorized personnel or criminal watch lists, after which a SentiVeillance-based application can immediately trigger a report or alert for recognized (or un-recognized) faces.
Updates in SentiVeillance SDK 4.0 include:
- Faster face detection and tracking,
- Improved facial recognition,
- Enhanced pedestrian detection and motion tracing,
- Improved tracking of moving objects with adjustable motion sensitivity,
- Better motion handling for objects or people that are very close to the camera and take up more than 20% of the field of view,
- Camera calibration for ground plane definition,
- User pre-defined object classification based on tracked object area and speed.
As in the previous version, zones can be defined as patches (when an object enters a zone) or tripwires (when an object crosses a line in a forbidden direction). The new SDK can also determine the speed of moving objects and classify them accordingly (for example, tracking speed in a restricted area). SentiVeillance provides facial feature determination and age estimation by face as well as detection of emotions, smiles, closed-eyes, open-mouths, clear glasses and dark glasses.
The SentiVeillance SDK supports multiple cameras and video sources, enabling one computer to process images from several cameras. It also provides connections between SentiVeillance units deployed on different computers, synchronizing the databases so that the system works as a whole within the logical network. It enables real-time face detection, extraction and matching by providing parallelization of facial recognition functions for improved performance on multi-core, multi-processor systems.
The SentiVeillance 4.0 SDK is available through Neurotechnology or from distributors worldwide.