by Ben Perlmutter

The Future of Motion Detection: A Discussion with Joyce Reitman and Howard Kline



During Joyce Reitman and Howard Kline’s talk at #LetsGetSmart18 at ICSC RECon, Howard interviewed Joyce about Motionloft, her company that uses proprietary software to track people and vehicles in real time.

Joyce Reitman is the CEO of Motionloft, which is based in San Francisco and is owned by businessman and investor Mark Cuban. Motionloft’s software services all of North America and just started operating in Japan. It’s used by global brands, commercial property owners, and smart cities.

 To track vehicles and people in real time, Motionloft places cameras in a location where they can see everything that’s happening in a given area. Software attached to the cameras then use artificial intelligence and machine learning to monitor how people and vehicles are moving. The data gathered from the software is then delivered to a dashboard for the customers’ personal use.

Howard began their interview by asking Joyce what is exciting about Motionloft. Joyce responded that Motionloft is special because the software collects data that was never able to be collected before. Before people would go to a street corner and count people and cars. Motionloft, in contrast, can collect orders of magnitude more data using its software based on artificial intelligence (AI) and machine learning.

To relate the high-tech alchemy of Motionloft to more relatable, simple technology, Howard then asked if Motionloft is “kinda like those cables that go across the street to track the number of cars that pass. Joyce replied, yes it “kinda” is, but much, much better. Motionloft can “take in the scene,” like a brain. Using its cameras, the Motionloft software creates a stereoscopic image of the environment. This means that the cameras  can percieve depth perception, like the human eye. The software also has “path track,” which follows where people and vehicles are going on a map. Additionally, there is an “engage” feature that notes where people stop to do something.

Next, Howard and Joyce discussed some of the future developments of AI that Motionloft could be involved in. Currently, Motionloft mainly does motion detection and identification of demographic info, like if a person is a man or woman, or that person’s age bracket. In the future, Motionloft’s AI platforms could expand to look at people’s emotional states as well.

They also talked about another future development of AI,“predictives,” which is using data to predict outcomes. Joyce provided an example of some research on predictives that Motionloft is currently carrying out. They are currently studying the effects of rainy versus sunny weather to see how people walk on the street. Using the observed data, they will be able to predict how people will behave in future weather conditions. While it’s common knowledge that less people walk on the street when it’s rainy, Joyce explained, “we don’t know how many [people], or at what time…we don’t really have a finer tuned sense of that.” Motionloft’s research will help create that detailed picture, which could provide valuable information for marketers.

Howard and Joyce concluded their discussion talking about some of the more uncomfortable aspects of this cutting-edge motion detection technology. Howard was concerned about his privacy in a world where one is constantly monitored. Many others are also uncomfortable about this 21st-century “hot topic.” Joyce assuaged his fears, explaining that while Motionloft uses cameras, it doesn’t save any of the video its technology records.  Rather, the videos are deleted immediately after the data is extracted from them, a process that only takes milliseconds. “Motionloft is interested in where you are, not who you are,” Joyce said.

After the talk concluded there was Q and A with interesting discussion about topics including Motionloft investor Mark Cuban, the company’s clientele which ranges from gas stations to banks, and a request to monitor alligators at a Disney theme park.

You can watch Joyce and Howard’s discussion here: