Home > Photography > Zensors app uses your old smartphone and…

Zensors app uses your old smartphone and crowdsourcing for smart surveillance

Want to know how many times your cat jumps on the counter when you’re not around? Or, maybe you want to quantify the distraction you experience at work by counting the number of people who walk by your cubicle. Now there’s an app for that. Zensors, created by a research group at Carnegie Mellon University in participation with the Yahoo InMind Project, uses your smartphone’s camera to capture data, and then uses artificial intelligence and crowdsourced input to analyze it.

The app, currently in beta, uses your phone’s camera to monitor the action. You can set up an area on the phone’s touchscreen for the application to pay particular attention to. A Zensors demo used a parking lot as an example, where the user defined a certain parking space, or group of spaces, to watch for a period of time. Data is then analyzed by set parameters, such as how many cars park in a certain spot over a period of time.

You can use your current phone, though the group behind Zensors suggests that this is a practical use for an old smartphone that you no longer use. You can set it up to be a dedicated data collector; if you use your current device, you may interrupt your data collection activities when you receive a text, or have to walk away from the scene.

zensors

The data is crowdsourced, because certain questions are not easily answered by the simple action of an object on the screen. One of the creators, Chris Harrison, assistant professor of human-computer interaction at Carnegie Mellon, writes on his profile page that certain sensors, such as a simple open/close door sensor, does not answer the question of whether someone’s children are home from school (although you might want pick who you want to crowdsource on that query). Another example is which patrons at a restaurant need their drinks refreshed. There are business and practical applications for the Zensors app.

For the crowdsourced monitoring, the job could be done by an outsourced staff that watches for activity. At the same time, the app’s algorithms are learning from the humans; over time, some of those human-based activities could become automated.

Crowdsourcing is sometimes necessary. Harrison notes that advances in sensing, computer vision, and machine learning are quite advanced, but not quite there yet to address the queries people might use the app for.

Zensors was built for use with unused smartphones, but will also work with a Wi-Fi camera such as a Dropcam. The site says users can set alerts by email or text, and adapt the app using the available API.