TRASSIR Neuro Left Object Detector
Overview
TRASSIR Neuro Left Object Detector
The module is designed to detect lost and forgotten objects left in the camera's field of view. The detector is useful in areas where security is especially important, for example, airports, train stations, in the "Safe City" on street cameras, etc.
Neuro Left Object Detector is based on an object and people detector. Unlike the classical model, it is not adversely affected by the background and works well in difficult conditions. The detector is based on two grids and tracking. The first network detects people and various classes of objects: suitcases, backpacks, boxes, bags, and duffles. The second uses frames of objects recognized by the first network to create their unique features. Tracking is the main task of the object detector. Works using the unique features of individual objects. Having carried out basic heuristics over the tracks, it determines the object's owner (by the nature of movement and proximity to the object), and whether he left the frame or moved away a certain distance.
Functionality:
The module is designed to detect lost and forgotten objects left in the camera's field of view. The detector is useful in areas where security is especially important, for example, airports, train stations, in the "Safe City" on street cameras, etc.
Neuro Left Object Detector is based on an object and people detector. Unlike the classical model, it is not adversely affected by the background and works well in difficult conditions. The detector is based on two grids and tracking. The first network detects people and various classes of objects: suitcases, backpacks, boxes, bags, and duffles. The second uses frames of objects recognized by the first network to create their unique features. Tracking is the main task of the object detector. Works using the unique features of individual objects. Having carried out basic heuristics over the tracks, it determines the object's owner (by the nature of movement and proximity to the object), and whether he left the frame or moved away a certain distance.
Functionality:
- The quality of the network, which affects speed and accuracy. The detector works better on complex scenes as quality increases. However, this also increases the load on the server.
- Number of frames per second collected for the detector to operate. More frames will result in a lower number of false positives, but a higher load on the video card.
- Minimum and maximum size of the object to be detected.
Other Products
