Over the past decade, digital cameras have become ubiquitous in modern life — embedded in smartphones, smart eyewear, security surveillance systems, autonomous vehicles and facial recognition technologies. And as the sheer amount of image data being captured has grown, so have concerns about privacy protection.

But what if there was a way to take pictures that instantly capture only the objects of relevance in a frame while simultaneously blotting out unnecessary or potentially sensitive details, without the need for any editing, encryption or other digital post-processing work?

In a paper recently published in the journal eLight, a team of researchers from the UCLA Samueli School of Engineering describe an artificial intelligence–informed smart camera that does just that.

Their device consists of several transmissive, 3D-printed surfaces (see image above), each composed of tens of thousands of diffractive features at the scale of the wavelength of light. The team used deep learning–based training to design and assemble the successive layers — training, in this case, that instructed the layers to only recognize the handwritten number “2” while effectively ignoring other numbers.  

As the camera lens captures its field of view, the layers let through light matching the number 2, creating high-quality images but diffract light from other numbers, like 5 and 9, instead forming random and low-intensity patterns that became noise in the background of the image.

Although the technology is still in its early stages, its potential to erase sensitive information at the point of capture could eliminate the risk of hacking or the exposure of original, unedited versions of images, especially those stored in the cloud. At the same time, it removes the need for any after-the-fact processing, which is data-intensive and consumes more power.

“We’d like to design a robust, power-efficient and secure smart camera that can safeguard privacy by controlling what a camera lens sees and captures before one clicks the shutter,” said lead researcher Aydogan Ozcan, the Volgenau Professor for Engineering Innovation at UCLA Engineering.

Read the full news release on the UCLA Samueli School of Engineering website.