Recognising what is in the image.

Once the objects are detected, Neural Object moves on to segmentation. This process isolates each individual object within the image, allowing a detailed analysis of each one. It improves Neural Object's ability to understand and analyse the visual environment.

The product also lets you set control points. Up to 4,096 control points can be set per processing cell and up to 4,096 cells per event, which works out at more than 16 million points. That gives users a high level of control and customisation over how and where object analysis is carried out.

Once segmented, Neural Object classifies the objects into different categories. This classification stage gives an organised, clear understanding of the different types of object present in the field of view. On top of that, Neural Object's counting function allows a detailed quantitative analysis of the environment, counting the number of objects in each category.

It works with most IP cameras on the market; exactly which ones depends on the installation and is checked before anything is promised. This broad compatibility makes Neural Object an adaptable, versatile solution that can be integrated into a variety of existing systems.

Neural Object stands out for its robustness and reliability in different lighting conditions and environments. Changes in light do not affect it, and it works perfectly indoors and outdoors, by day or by night.

Neural Object's effectiveness is absolute, with a precision of 98-99.9% in detecting people, even at entrances with massive crowds. That works out at a real margin of error of 0.1-1%, guaranteeing maximum precision in every environment.

Thanks to its ability to adapt and learn over time, Neural Object keeps up with any change or variation in the visual environment. As an integral part of the Infinity Neural solutions, Neural Object is an invaluable tool, providing high-quality visual analysis that stays ahead of computer vision technology.