Hardware requirements for neural analytics operation

Neural analytics is a detection tool (see General information about detection tools and their sub-tools), the algorithms of which use neural networks.

Neural network is a mathematical model, as well as its software or hardware content, built on the principle of organization and functioning of biological neural networks (networks of neurons of a living organism). Within the context of the described software environment, neural networks are file models that have the .ann extension that you can use in the detection tools operation.

The table lists the detection tools that use neural networks in their operation:

Detection tool that uses neural networkAction of the neural network
Face Detection TVDetection of faces
Face Detection VA
Face Detection VL
License Plate Recognition BRSDetection of vehicle license plates
License Plate Recognition IV
License Plate Recognition RR
License Plate Recognition RR—Search In Archive
License Plate Recognition RR—Parking
License Plate Recognition VT
Vehicle Recognition XR
Vehicle Recognition RRDetection of vehicle attributes
Object Tracker with a neural network filterProcesses tracker results and filters out objects of no interest in a complex video image (foliage, glare, and so on)
Object Tracker VLDetection of people in the frame
Stopped object detectorDetection of stopped objects in the frame
Abandoned Objects Detection VIDetection of abandoned objects or objects that disappeared in the frame
Abandoned Objects Detection (Street) VI
NeurotrackerDetection of the position of only the required objects in the frame
NeurocounterCounts the number of objects in a specified area
Fire DetectionDetection of fire in the frame
Smoke DetectionDetection of smoke in the frame
Object Presence DetectionNeural network is trained to detect a specific object in the frame

Pose Detection and its sub-tools:

Identifies each person's "skeleton" and detects poses that can represent a security threat
Equipment detection (PPE)

Detection of necessary equipment and personal protective equipment in the frame. Segmenting and classification neural networks are used for the operation of the Equipment detection (PPE) (see Functions of the Equipment detection (PPE))

PPE detection VL
Person-Based Privacy Masking

Segmenting neural network breaks down the human body into zones

Privacy MaskingDetection and masking of moving and stationary objects
Water Level DetectionIncreases the accuracy of water level detection in difficult conditions. For example, in scenes with clear water
Audio Analytics Detection IVRecognizes different sounds: from a baby cry to the sound of a gunshot.
Crowd Estimation VACounts crowds of people in a specified area
Barcode DetectionDetection of a specific type of barcode in the frame