Skip to main content

Key Concepts

This section covers a few basic concepts that will help you use EyesOnIt effectively.

Detection Regions​

Detection regions give you an easy way to tell EyesOnIt what part of your image to focus on. Click here to learn all about detection regions.

Object Descriptions​

Object descriptions are the text you enter to tell EyesOnIt what you want to detect. Click here to read all about them.

Confidence Levels​

EyesOnIt uses confidence levels to tell you if it has detected the objects described by your descriptions. This page provides everything you need to know about confidence levels.

Alerts​

EyesOnIt brings together object descriptions, confidence levels and detection regions to provide an alert when it detects the object you care about. This page explains the power of EyesOnIt alerts.

Advanced Concepts

This section describes a few advanced topics that will help you optimize EyesOnIt for scaled deployments with many streams.

Scalability​

Scalability is a key consideration for any real-world deployment. EyesOnIt includes several features to reduce deployment cost by maximizing the use of your hardware. Click here to read the scalability overview and dive into the details.

Frame Rate​

The frame rate setting is critical to optimizing the use of your CPU and GPU. Click here to read about tuning your detection settings to process one hundred streams or more on a single server by controlling the frame rate of each stream.

Motion Detection​

Motion detection allows you to optimize the use of your CPU or GPU by only running the EyesOnIt model when there is activity to analyze. Click here to learn how motion detection can help you scale your deployment.

Common Object Detection​

EyesOnIt can use a pre-processing step to detect common objects including people, vehicles, and bags. This step helps EyesOnIt focus the full model on the area of greatest interest within your detection regions. Click here to learn more about common object detection.