Mirasys VMS V9.10 RELEASE NOTES
- Mirasys Support
- 11 minutes ago
- 5 min read
Release date: 8 Jul 2026
HIGHLIGHTS FROM RELEASE NOTES
Enhancements
Recognition functionalities enhancements:
Libraries update
The AI model inference and video decompression libraries have been updated to newer versions. These updates improve speed and reduce resource consumption for face, license plate, and object recognition.
Installer changes
Installation of the Mirasys License Plate Recognition, Face Recognition, and Object Recognition services requires the installation of the following NVIDIA libraries for Windows:
CUDA Toolkit 12.9 Update 1: https://developer.nvidia.com/cuda-12-9-1-download-archive
cuDNN 9.20.0: https://developer.nvidia.com/cudnn-9-20-0-download-archive
TensorRT 10.16.0.72 GA for Windows 10, 11, Server 2022, and CUDA 12.0 to 12.9 ZIP Package: https://developer.nvidia.com/tensorrt/download/10x
The installers detect NVIDIA library installers in the ‘NVIDIA’ subfolder and install them automatically if present. Please find the detailed installation steps in the installation instructions.
Face recognition improvements
Face Recognition has undergone significant changes to improve face detection and identification accuracy without affecting performance. Integrating face tracking with recognition eliminates multiple recognition events for the same face, reducing false positives and duplicate detections, enabling users to focus on genuine recognition events.
The system now detects up to 20 faces simultaneously from the video stream.
Object recognition improvements
Object Recognition has been enhanced to detect persons and vehicles, as well as their attributes, more accurately, resulting in fewer false positive detections.
In the camera Object Recognition settings, the user can select which object recognitions are performed for the camera stream. The user can select whether vehicles or persons are detected, and if the person similarity data is being calculated for the person similarity search. This allows for optimising the Object Recognition resource usage and reducing the amount of detection data by selecting only those object recognitions that are truly needed.
Smart Recognition and Smart Search improvements
The face recognition event in the Spotter Smart Recognition and Smart Search event list displays both the matched identity's default image and the matching face image. This lets users identify which identity matches the detected face and the exact face image matched.
Filters in Smart Recognition and search criteria in Smart Search support free text search, enabling quick selection when filter values or criteria lists are long.
List management improvements
Identity Image Refactoring
The usage of identity images in the List Management has undergone a major refactoring. The number of identity images for an identity is limited to three images. The user can define which identity images are used with the face recognition. The identity images used with the face recognition need to meet the new requirements: the image must contain a big enough face, and the quality of the image needs to be fit for the face recognition and face matching. This is ensured by calculating the face image quality value when an identity image is added to be used with face recognition. The identity images that do not meet the requirements cannot be used for face recognition.
System and Spotter Smart List Management Settings
The identity and the identity list settings in the System Manager List management settings and in Spotter Smart List Management have been changed to show all identity images for each identity with the identity face image quality. The quality value is shown for the identity images that are selected to be used with face recognition. The identity lists can be sorted by identity image using face image quality.
Identity Import and Export Updates
The identity and identity list import and export have been updated to support the new identity image parameters. By default, identity import from the older system will try to set all identity images to be used with face recognition. The import progress indicator and the import result will show details of the import success and failures. The identity export can be used to export only the identities that have identity images with a low quality value. This allows easier bulk update for the low-quality identity images.
Peer-to-Peer review
A new Peer-to-Peer functionality allows Spotter operators to share their selected view with non-security workers.
By using the new Spotter Viewer application, the non-security worker can see the operator's view as the operator sees it, allowing the reviewing of shared footage without further access to the security system and its stored data. Install the Spotter Viewer using its separate installer.
Mirasys VMS system supports several simultaneous viewing stations to which the Spotter operators can share their view. The view-sharing rights per viewing station can be defined at the user group level, and the view-sharing activities are stored in the audit trail.
Spotter PTZ and joystick control improvements:
Support for Hanwha SPC-7000
Support for the Hanwha SPC-7000 joystick was added to the Spotter application.
PTZ speed controls for joystick
In the Spotter input device layout settings, users can adjust the PTZ speed for pan, tilt, and zoom independently. This enables precise control of PTZ movement at the joystick level.
Maximize and restore the view in PTZ control mode
Double-clicking the mouse to maximize or restore the view in the Spotter PTZ camera control has been disabled. This change ensures the point-to-move PTZ control functions correctly without triggering view maximization. When PTZ control is active, use the maximize and restore buttons to adjust the camera view, as you would when PTZ control is inactive.
Spotter export functionality improvements:
Dual export
Exporting to two different locations with different export parameters is now possible in the Spotter export by using dual exporting. This allows the creation of exports for different purposes, for example, one for archive purposes and another for authorities. The dual exporting can be used with clip, storyboard, storyboard draft, alarm, and still image export.
To enable dual exporting, the Spotter role for it has to be enabled first in the System Manager user group settings. When the user has the right to use dual exporting, the default export settings for the dual export can be adjusted in the Spotter settings. The dual export parameters can also be adjusted in the export dialog.
Confirmation on the storyboard clearing
Click the Clear button in the storyboard panel to clear its content. To prevent accidental clearing, the system requests user confirmation and clears the storyboard only after confirmation.
Allow more than 8 channels to export.
Users can add up to 32 streams to a single SEF export in the Spotter application, allowing more streams to be saved together. Exporting multiple streams in one file provides an efficient method for archiving data. If other formats are selected, the export dialog informs users of stream count limits, exporting only the first streams supported by the format.
You can download the full release notes as a PDF file here: 9.10 - OneDrive
Download the new version directly on the Extranet or here. In case you have questions, please contact Mirasys Technical Support.
If you are interested in acquiring Mirasys solutions, get in touch with a Mirasys expert: sales@mirasys.com
Comments