Below are answers to some of the most frequently asked questions about the IntelliSee risk mitigation platform. If you have a question not covered below, or would like to speak to an IntelliSee representative, please Contact Us
IntelliSee is an AI risk-mitigation platform that leverages existing surveillance cameras to autonomously detect risks and threats to your organization. It keeps your organization safe by then alerting your systems and people, in real-time, about these risks so they can act to prevent incidents and tragedies.
The vast majority of surveillance is unmonitored and mainly used as a visual deterrent or for forensics after an incident has occurred. When monitored, overtaxed security staff are charged with addressing threats while also finding them. Live monitoring is a difficult and error prone task given people either zone out (“screen fatigue” begins in less than 15 minutes) or zone in/get tunnel vision and can miss other potential risks. IntelliSee automates this mundane task so an organization’s surveillance investment can evolve from a reactive, passive tool to a proactive risk mitigation tool allowing security personnel to react and respond quicker.
IntelliSee is an easy add-on to your existing surveillance system. Rather then requiring an organization to upgrade their existing cameras/surveillance system, IntelliSee makes your existing surveillance “smarter” by adding an affordable, non-disruptive layer to your existing safety and security strategy.
IntelliSee uses advanced artificial intelligence (AI) called deep-learning computer vision and powerful computing (GPU’s) to analyze the pixels of surveillance footage to detect visual risks and threats. IntelliSee trains its advanced AI algorithms on surveillance footage containing a wide variety of risks, backgrounds/environments, and conditions so that it can properly identify what the camera “sees”. To date, our incident database contains over 30 million frames of “bad things” and it continues to grow.
IntelliSee is a platform designed to autonomously detect multiple risks at the same time. Risks range from common slip & fall hazards such as spills and puddles to the potentially catastrophic presence of a drawn gun. IntelliSee also can detect the presence of a person trespassing or the absence of a person in certain circumstances so you will know if key personnel are not where they should be. IntelliSee’s goal is to cover as much of the risk matrix (probability vs. severity) as possible so new risks will continually be added over time.
IntelliSee is fully customizable through its web portal so users can set which cameras to monitor, when and for what risks. For example, users could set a trespassing detection during the night when people aren’t expected to be in an area.
IntelliSee can even alert users if a surveillance camera has been disconnected or quit working.
IntelliSee’s connects to the same network as your surveillance cameras and monitors the live data feeds via a simple ethernet cord. Depending on your surveillance approach, this generally happens through a network switch.
Setup is non-disruptive and involves IntelliSee securely scanning your network for your cameras’ IP addresses and then choosing which camera(s) to monitor and setting each one up with the desired detection and alert parameters within the IntelliSee portal. During this process, administrators will need to have the IP address, make/model, and login credentials for each camera to be monitored by IntelliSee.
IntelliSee works with any camera between 480p and 4K resolution.
Because IntelliSee taps into existing surveillance cameras’ live feed, it technically does not require a video management system (VMS). However, VMS’ are often an important piece of an organization’s overall safety and security strategy (e.g. to recover video footage surrounding an event). As such, IntelliSee is designed to seamlessly integrate with VMS and other critical systems (e.g. communication systems, access controls, etc.).
Once a threat is detected, IntelliSee instantly verifies the threat and sends alerts based on a user’s specific parameters. Because most organizations have standard operating procedures (SOP’s) and other existing safety and security systems, IntelliSee’s alerts can and should be integrated with other existing systems. These are referred to as “system-to-system” alerts.
For redundancy, IntelliSee email alerts are also recommended. IntelliSee can also send text and phone call alerts, but alerting via these methods also relies on other system providers (e.g. your cellular provider) which can delay receipt of the alert. For this reason we recommend system-to-system alerting paired with email for time sensitive alerts. Texts can be used for alerts that are not time sensitive.
Because each organization’s approach to safety and security is different, it’s important to work with your IntelliSee representative on what integrations will be best for your organization. Some examples of common system-to-system integrations include:
- Integrating IntelliSee trespass alerts with control room/video monitoring systems (VMS) to draw the security team’s attention by displaying a viewing window for the camera detecting the trespasser.
- Integrating IntelliSee fallen person alerts with VMS/DVR systems to bookmark footage (for future retrieval) and with communication systems to alert staff that someone has fallen.
- Integrating IntelliSee weapon alerts with dispatch to initiate a lockdown safety protocol and/or access control systems to lock doors if someone is approaching the building with a gun.
- Integrating with a 2-way radio system to communicate a trespasser or gun situation.
Although IntelliSee can detect risks across the risk spectrum, we can’t detect everything yet. There are two requirements for risk detection:
- The risk has to be visible. For example, a weapon stored in a guitar case will look like a guitar case to the surveillance camera and IntelliSee.
- IntelliSee’s deep-learning neural network has to include the risk to be detected. We have to first tell the AI what to look for, such as a spill or a gun before it can be detected. The number of risks “learned” and the amount of incident footage IntelliSee learns from continually increases. If you have a risk you would like IntelliSee to learn please contact us so we can discuss.
The best rule of thumb is if it’s viewable and knowable by human eyes (without any pre-information about what is being looked for), IntelliSee will be able to detect the risk if our deep-learning neural network has learned about the risk.
As a deep-learning AI platform, IntelliSee continually improves as more footage around various incidents and various environments/backgrounds is added. Each release improves the IntelliSee platform so all users benefit from its continual learning. Unlike a car that degrades the more miles it’s driven, IntelliSee improves the more it is used by everyone.
There are three basic rules to describe the limitations of computer vision based AI:
- IntelliSee can only see what the camera can see (e.g. if a camera is obstructed IntelliSee can’t see the threat either.)
- Artificial Intelligence is not the same as human intelligence (e.g. we don’t know if someone is lying on the floor to stretch or has fallen and needs help.)
- Angles, distances and lighting matter. (e.g. if a small object is very far away we can’t detect it.)
IntelliSee uses computer vision deep-learning AI which is more powerful than machine learning AI. With machine learning, a human programmer still determines what attributes to analyze. Deep-learning operates similar to the human brain with deep neural networks so it is “self-learning.”
False positives (detecting something that is not actually a risk) and false negatives (missing a risk) can occur with AI. While no system is foolproof, IntelliSee uses advanced deep-learning computer vision AI and is designed to validate risks in real-time. This step helps to minimize false positives and false negatives. The IntelliSee user-interface also gives users several options to further manage these potential situations.
IntelliSee takes your privacy and security seriously. IntelliSee has elected not to implement facial recognition due to privacy concerns, an evolving regulatory environment, and the potential for misuse. Making organizations safer does not require facial recognition. IntelliSee retains no personal identifying information (PII) and does not store video footage.
IntelliSee uses your organization’s footage in order to optimize its performance at your site. IntelliSee requests that each customer (or their Systems Integrator) upload 10 minutes of footage from each camera being monitored during an active time period to learn the environment and background. Many organizations also send historical incident footage to make IntelliSee even smarter for their environment. IntelliSee does not learn on site (the computational power required is too significant) but as you upload additional video footage around incidents, IntelliSee is trained and new releases are deployed to allow your organization to benefit from your learning as well as the learning from all other IntelliSee installations.
The IntelliSee appliance is located inside the client’s firewall and only communicates outside that firewall to update configurations and send email, text or phone alerts. IntelliSee has a secure configuration process, and we do regular 3rd party penetration testing.
Bandwidth varies by camera manufacturer and resolution, but a good estimate is less than 1 MB/s per camera. To minimize bandwidth the user would have to reduce the resolution of the camera independently of the IntelliSee application.