SmartVision 6.3 has been released. This update focuses on improving video analytics performance, enhancing GPU-based neural network detectors, and making the system easier to configure for different computer setups.
In SmartVision, line-crossing detectors have been optimized for GPU processing. This is especially important for video surveillance systems with multiple cameras, where events need to be detected quickly while reducing delays in video stream processing.
Gender and age detectors have also been accelerated when using a graphics processor. Moving these tasks to the GPU allows neural network models to run more efficiently, since graphics cards are much better suited for the parallel calculations used in modern video analytics.
A new setting has been added to SmartVision that allows users to specify how many frames per second are sent to recognition detectors. This makes it possible to fine-tune the balance between speed, accuracy, and computer load. For example, a powerful workstation can process more frames for improved analysis, while a standard PC can reduce the workload and maintain stable performance.
Version 6.3 also improves the quick message window, Alerts. It is now more convenient for monitoring events, notifications, and important system triggers. In addition, the interface has been refined, and translations have been updated for multiple languages.
Special attention has been given to checking the components required for hardware acceleration. SmartVision now performs additional checks for CUDA and cuDNN installation, and also verifies the availability of DirectX and DXCore. If the required modules are not installed or are unavailable, the system will continue to run on the CPU. This keeps SmartVision compatible with regular computers, even when GPU acceleration is not available.
CUDA, cuDNN, DirectX, and DXCore: Why They Matter
CUDA is an NVIDIA technology that allows a graphics card to be used not only for displaying images, but also for high-performance computing. This is especially important for neural networks. Object detection, face recognition, license plate recognition, age and gender detection, and other AI-based tasks require a large number of repeated mathematical operations. A GPU can perform these operations in parallel, which can significantly reduce the load on the CPU when the system is properly configured.
cuDNN is a specialized NVIDIA library for deep neural networks. It includes optimized operations commonly used in neural network models, such as convolutions, matrix calculations, normalization, pooling, and other functions. In simple terms, CUDA gives the system access to the computing power of the graphics card, while cuDNN helps neural networks use that power efficiently.
DirectX is a set of Microsoft technologies used in Windows for graphics, multimedia, and video acceleration. In video analytics, DirectX can be important for proper interaction with the graphics subsystem, drivers, and GPU capabilities.
DXCore is a Windows component related to detecting and listing available graphics and compute adapters. It helps applications understand which video devices are available in the system. If DXCore is missing or damaged, the software may not detect the GPU correctly or may be unable to access the required hardware acceleration features.
As a result, SmartVision can now better adapt to the capabilities of each computer. If an NVIDIA GPU, CUDA, cuDNN, DirectX, and DXCore are installed and available, the system can use hardware acceleration for faster neural network detection. If these components are not available, SmartVision will continue running on the CPU without requiring complex manual configuration.