A diffractive camera that can selectively forget the things it captures.

Privacy is an obvious concern now that everything from smartphones at smart watches same smart glasses has built-in cameras. Banning hidden cameras will never happen, and digitally altering footage for privacy is a real pain. So, instead, the UCLA researchers are working on a radical new type of camera who can selectively capture or ignore specific objects in the frame even before they are checked in.

If you’ve ever seen an investigative news program protect a source’s identity by blurring or pixelating their facial features, you’re already familiar with one of the many methods we already use to maintain confidentiality. Other approaches include encryption of sensitive media or more advanced processing techniques that digitally erase part of a photo using tools like Photoshop. There are too automated algorithms, which services like Google Maps use to blur faces and license plates in billions of photos.

However, these are all post-processing methods that occur after capturing and storing a digital image. The original unprocessed images potentially containing private data still exist and could still being exposed – something we’ve seen happen time and time again – which is why UCLA researchers wanted to address privacy issues at the source: when light enters a camera, but before it reaches the image sensor.

Camera makers could potentially release firmware updates with AI-powered tools that, for example, could be used to selectively erase specific people from a photo. But that requires a level of processing power that even a high-end digital camera might not have, so UCLA researchers solved the problem optically, using a technique they call “the diffractive computing”, as detailed in a recently published article.

A diffractive camera capable of selectively erasing what it sees.

Even if you know photography well, this camera takes a radically different approach to capturing images. The researchers started with a desired object they wanted to record – in this case, some very simple black-and-white, hand-written numbers – and used it to train a design tool based on deep learning. which generates a series of diffractive images. layers that can be 3D printed and assembled in series to create a “computer imager” that sits in front of an “exit plane,” where the final image is captured.

Each layer has tens of thousands of microscopic diffractive features that are specifically designed to allow light that corresponds to desired objects to pass through unaffected, while light from other objects is diffracted and optically erased into insane patterns and low intensity that look like chance. noise. This means the image that is actually captured at the end cannot be reverse engineered to extrapolate what was removed.

As you can probably imagine, the practical applications of this radically different approach to photography are incredibly limited at the moment. You won’t see a ‘do not capture Uncle Bill’ feature added to the iPhone’s camera app anytime soon.. But research offers impressive advantages over current techniques. Not only does “image processing” literally occur at the speed of light, as it is all optical and analog, but the design of the diffractive layers could also introduce optical encryption, hiding details in a photo that can only be revealed using a decryption key that shows how the original image can be recovered.

Previous

Have a dizzying summer adventure with this 4K camera drone for less than $100

Next

The Best Camera Bags for Landscape Photographers This Fall

Check Also