Irisity continue building its IP and innovation with three new patent approvals.
System and Method for use in training machine learning utilities – training deep learning models for use in AI requires large datasets of annotated (labeled) data. This annotation task is mostly done by hand or using semi-automated tools. The patent granted describes a method for automatically annotating data for use in machine learning models.
System and Method for use in Geo Spatial Registration – some applications of video surveillance require taking the position of an object captured by a security camera and translating it to geo-spatial coordinates so that the object can be represented on a GIS map. To perform this task, the camera needs to be registered on the GIS map, a very time and labor consuming manual task. This patent describes a method for automatically performing the registration by having a person walking in the field of view of the camera and using a special app installed on that person’s phone. The patent allows customers with large camera systems (for example cities) to register all their cameras quickly and with a minimal amount of effort.
System and Method for use in Object Detection from Video Stream – using AI to process high resolution live video streams is a computationally expensive task and as a result building large networks of intelligent cameras can be computationally prohibitive. This patent describes a computationally efficient method for applying AI to high resolution streams with drastically reduced computational resources and while maintaining high accuracy.
The patents above have been granted in Israel establishing priority date, with national phase examination in additional countries to follow. These three patents complement our 9 patent families with in total 32 patents and applications, of which 11 applications and 21 granted patents.
At Irisity the innovation is flourishing, making us more responsive and agile in our development of new stellar solutions for an ever-growing market demand.