Use artificial intelligence to enhance your digital images




By now, we all have accumulated thousands of digital images over the past two decades or more. And many of us have spent the pandemic years catching up on scanning even older images from the days of film cameras.

There’s a suite of tools that can provide some amazing enhancements to these images, taking advantage of artificial intelligence models to improve the look of your still—and video—images. Topaz Labs ( offers a set of three tools, DeNoise AI, Gigapixel AI, and Sharpen AI, that let you work on different issues that come up in digital images.

DeNoise AI lets you reduce the amount of digital noise that is inherent in the digital photo format. The Gigapixel AI program allows you to increase the size of an image with a significant reduction in the digital pixilation that happens when you enlarge a low-resolution image. And Sharpen AI improves the focus of images that have blur or are simply out of focus.

There is also a combination program called Photo AI that combines all these elements in a single program.

All of the programs also work as plugins in Adobe Photoshop, so it’s very convenient to add them to your post-processing photographic workflow. They understand the Camera RAW format that professional photographers use, and the programs automatically save the treated photograph as a new file by appending information to the file name so you remember which program worked on it.

You get mixed results when you use these tools on scanned images from film photos. I tried enhancing a few Kodachrome transparencies my dad made at the Garden State (PNC Bank) Arts Center in Monmouth County back in the early 1970s. Images of some performers got a lot sharper, but it didn’t do much on others. I’m sure it has something to do with the quality of the scan it’s working on.

You will get much better results using these programs on images that originally began as digital photos. I tried the Sharpen AI program on a couple of night photos I shot in Israel using long exposures while hand-held (hint: use a tripod or brace on a surface). Sharpen AI actually brought out more detail in parts of the image that were not moving but got blurred by the long exposure. A night image I made of the Shuk HaCarmel in Tel Aviv shows a lot more detail in the buildings and people on the street after processing.

With some more work and a lot of trial-and-error, you can surely recover some images you thought might be rejects.

If you work in video, Topaz offers Video AI to deal with video upscaling, motion, and shake stabilization. You can dramatically improve your video using automatic presets that call on AI models in the cloud to fix your recordings, or you can tweak the models to achieve the effects you need.

The downside of these programs for casual users is that they do expect a lot of horsepower in your computer. Topaz indicates you need Windows 10 or 11 64-bit, and on Apple, you need Mojave 10.14 or above. You can use the programs with 8gb of RAM but they recommend 16gb or more. And you need a high-performance graphics processor like Nvidia GTX 740 or AMD Radeon 5870.

Even with a high-performance computer, expect the video program to take quite a while to enhance a long video. I probably wouldn’t recommend running a two hour wedding video through the program but if you have shorter clips of a few minutes each that you wish looked better, it will do a nice job on them—and you won’t need to wait overnight for the results.

The programs themselves are surprisingly affordable. You can get the Photo AI program— or a suite of the three standalone programs (Sharpen, Gigapixel, and DeNoise)—for $199. Keep your eyes out for periodic sales that include all four for one price. The Video AI program is $249 and occasionally is offered at a sale price.

Email and tell me how you process your digital images. Follow on Mastodon.

Leave a Reply

Your email address will not be published. Required fields are marked *