AI (artificial intelligence) is increasingly becoming a part of day-to-day life, even if it isn't always obvious. While the term may conjure images of sentient robots, in the tech world it mostly refers to using computers to automate tasks once performed by people. These tasks usually involve some kind of informed decision-making. Until recently, those decisions were the exclusive domain of humans, but advances in technology like machine vision and speech recognition are letting computers catch up quickly.
What is an AI Camera?
Features like voice activation technically qualify as AI, though the term "AI camera" usually refers to more sophisticated technology. AI cameras use computer software (usually based on neural networks and computer vision) to:
1. Identify photos.
2. Automatically retouch them.
3. Help you capture better photos in real time by evaluating the shot before you take it.
The software that powers these features was designed in large part to compensate for the ubiquity of smartphone cameras and their limitations. In general, smartphone cameras don't have features like interchangeable lenses or adjustable focal length. As a result, developers have been working on software designed to digitally mimic the kinds of effects a photographer might achieve with a DSLR and years of technical experience.
What is AI in Smartphones?
As the software, neural networks, and algorithms improve, AI cameras are beginning to offer professional ― or pseudo-professional ― looking editing effects to smartphone cameras. You'll also see this technology at work in less sophisticated point-and-shoot cameras and even professional grade cameras, where they subtly improve effect quality.
One impressive example of AI camera technology at work is facial re-touching. Here's a short guide to the transformation:
- Pre-AI. A few years ago, software was able to digitally remove imperfections and blemishes from a face by essentially blurring parts of the image slightly. This process erased some detail, making those small imperfections hard (or impossible) to see. The downside, of course, was that your photo lost details to achieve an effect. The texture that makes an image convincing and "real-looking" is the result of details, many of which were lost in the smoothing process. This loss of texture left a disconcerting plasticity to many people's faces in photos.
- Post-AI. Modern AI cameras are beginning to offer algorithms that allow you to remove imperfections and add a convincing artificial texture. This effect avoids the "plastic" look while still improving the re-touched image's quality for both amateurs and professionals.
AI Features in Smartphones
Another common application of AI camera technology helps amateur photographers create better shots. Some AI cameras can do the following:
1. Analyze the image data coming in through the camera lens.
2. Compare it to huge databases of related images using a neural network.
3. Offer suggestions about how to better frame the shot and adjust settings like exposure to make sure that the photo comes out well.
Still other AI cameras take it a step further and automatically adjust settings on the fly (such as moving automatically to a fast exposure time if the scene is identified as a sporting event).
Many applications are beginning to employ AI to help people save time as well as capture better photos. The Google Clips wearable camera can automatically delete low-quality images ― such as overexposed or blurry shots ― saving you from having to sort through the images and maximizing your storage space.
The same technologies that allow amateurs to improve their photos can save professional photographers a lot of time in the editing room. AI and algorithms can correctly identify and sort the best images or automatically apply edits that once had to be done by hand.