Always-on computer vision sensing: From science fiction to science reality

In the late 22^nd century, the teddy bear of tomorrow will be more than a lovable static plush – it will interact directly with your child, or even other members of your family, with always-on visual awareness, as seen in Steven Spielberg’s futuristic sci-fi drama, “AI: Artificial Intelligence.” That’s the future of toys, but we are closer than you may think.

Soon, toys of the future may read faces and distinguish family members, creating intelligent interactions.

Qualcomm Research has developed an affordable, always-on Computer Vision Module (CVM) that could soon turn science fiction into science reality. This state-of-the-art sensor module with embedded CV processing can act as an intelligent Interactivity Trigger that can be integrated into all the things you use on a daily basis ranging from IoT appliances to your children’s toys, allowing them to intelligently react to users.

For example, as your child interacts with her toy, CVM enhances the experience by triggering the toy to react in different ways based on the movements and gestures of your child even when your toy is “asleep.”

What’s best is that CVM mitigates the privacy sensitivities which can be magnified when it comes to our children. The CVM does not output images. The processing happens within the sensor, thus emitting only post-processed CV metadata, noting for example that there is a face looking at the toy. It’s simply a sensor that cues a device to “switch on” or intelligently react to a child (or other user or object).

Our CVM’s small size supports integration into numerous types of toys and IoT devices.

The movies that take place in the future, such as “AI: Artificial Intelligence,” paint an interesting vision of our world to come. At Qualcomm Research, we are working to turn this vision from science fiction to science reality and it’s closer than you think. Soon, every device and appliance that you own could be integrated with CVM, delivering a next-generation type of interactivity, user experience, and convenience we can all look forward to.

In brief, the Qualcomm Research CVM’s key benefits are:

• Power Efficiency: <2mW active, end-to-end power
• Affordability: We’ve created an entire module with embedded CV algorithms at an attractive cost for price-sensitive smart toys and smart home devices.
• Privacy: No image output – just computer vision metadata

Want to learn more? Be sure to check out our always-on computer vision module page. Also, sign up for our newsletter to receive the latest information about CVM.

Qualcomm Research is a division of Qualcomm Technologies, Inc.