How creatives are using AI to transform the way they work with images
AI is changing the design world. For many creatives, its most useful, interesting and exciting role will be as a tool, to help them with difficult and time-consuming processes.
Image-upscaling is a perfect example of this kind of process.
By using AI to make it better and faster, designers can open up new creative possibilities.
When it comes to content, working with older imagery is fascinating. But it can be very challenging on a practical level. Modern screens are unforgiving on low-res images, while old photos might be faded or damaged, rendering them much harder to use.
Tools such as neural filters in Photoshop can also be used to turn black and white pictures into vivid colour, bringing them to life in a sensitive and accurate way.
In the worlds of gaming and entertainment, NVIDIA’s GPU powered tech has long blazed a trail in upscaling moving images for high definition screens. This technology is built into GeForce-powered laptops and desktop PCs, where it plays a key role in delivering smoother, sharper gameplay.
But as AI has become prolific in almost every industry, it has also opened up new possibilities in creative workflows – both on the cloud and locally with NVIDIA’s AI-accelerating technology, powered by its GeForce RTX graphics cards.
Upscaling in action
Here are four examples which show how NVIDIA’s RTX technology can enhance AI-accelerated creative workflows by sharpening and upscaling low-resolution or blurry images.
Fox in the snow
You can see here the incredibly detailed results which appear through the upscaling process. This was achieved using ON1 Resize AI, upscaling the image 400%. From the fox’s fine whiskers to its life-like snout, mouth and eye, the animal becomes strikingly present. Even the snowflakes stand out more.
Crucially, the edge where its facial hair meets the background branches is sharply defined – an area where old-school upscaling often struggles.
Mountain scene
Here the upscaling pulls off three impressive visual improvements using Topaz Gigapixel AI and Photoshop.
Firstly, Photoshop’s AI neural filters recolour the scene, bringing out the greens of the foreground, the icy blue of the lake and the snowy mountains. Secondly subtle lighting and contrast adjustments highlight how the sunlight falls across – and reflects off – the snow and rocks.
Topaz Gigapixel AI then upscales the image, sharpening previously fuzzy areas like the trees and bushes, giving the entire scene more definition.
Mountain with birds
Generatively upscaled 400% through Topaz Gigapixel and enhanced with Topaz Bloom AI, this image shows the power of AI-assisted enhancement running on NVIDIA GeForce RTX technology. The RTX GPU accelerates the process, allowing subtle colour nuances and complex shades to appear naturally and render in moments, not minutes.
You can also see its sharpening in full effect – the before image is almost hard to look at because it’s so blurry, while every detail on the trees and the mountainside comes to life once it’s been upscaled.
The birds are interesting too – the tech knows that, because they’re in flight, they should appear somewhat blurry, but the new version is much cleaner to look at.
Woman in snow
This image shows how AI upscaling can enhance both close-up portraits as well as panoramic landscapes. Generatively upscaled 200% using Topaz Gigapixel AI with face recovery technology in Photoshop, we see the sharper details brought to life – the strands of the woman’s hair and the weave of her scarf.
But we also see the subtlety of its use of colour achieved through Photoshop’s AI neural filters – the different hues in the scarf, the red of her cold cheek and the restrained grey in her hair all add a sense of realism, and even storytelling, to this simple scene.
How AI-powered photo upscaling works
AI-powered upscaling works in a very different way to traditional upscaling software. That takes a low-res image and tries to drag it into a high-res format, with very mixed results, including blurry transitions and jagged edges.
In contrast, AI upscaling uses a deep learning model, accelerated by the hardware in a GPU, to predict and generate high-resolution details. This approach makes the upscaled image much more nuanced and realistic.
Nvidia’s GeForce RTX 50-series graphics cards set new standards in this space. Its specially developed tensor cores dramatically improve its AI capabilities, making it quicker and more powerful. For creatives, this means enhanced clarity and detail, helping bring their projects to life with next-level fidelity.
And thanks to this GPU-powered acceleration, users can upscale and enhance images without frustrating lags or delays that pull them out of their creative flow.
This technology already accelerates many of the leading design software apps, including Adobe Photoshop, Topaz Gigapixel AI and ON1 Resize AI, as demonstrated above. These platforms can leverage NVIDIA GPU power technology to balance effective image enhancement with accuracy and authenticity.
Upscaling images with AI requires a certain amount of imagination, but this needs to be done with a sense of responsibility, to create the most realistic image possible. This is especially important to preserve trust between designers and their audience.
The impact of AI upscaling
Visual communication is all about impact. A picture is worth a thousand words, as the old saying goes, and images have an ability to explain, illustrate and inspire that is both immediate, and emotional.
But to create the maximum impact, images need to be high quality. Or to put it another way, even the most incredible visual will be undermined if it’s blurry, pixelated or damaged.
So with AI-powered upscaling, you can ensure that every image is given the best chance to live up to its potential.
- If photos have degraded over time, or been damaged in storage, galleries, museums and archives can use upscaling software to breathe new life into them. This allows curators to bring these vital and vivid documents of the past into the present, and use them to educate and inform generations to come.
- Designers love working with archive material, but important parts of our visual culture sometimes survive as borderline-unusable low-res images. AI-powered acceleration in upscaling tools, designers can make any image high-quality, and it gives them the freedom to choose and use whichever visuals best tell their story.
- When time and money is tight – and let’s face it, when isn’t it? – teams can improve low-quality content, or fix production errors, without the need for costly reshoots and client delays.
Conclusion
At this moment in time, many designers are wary about generative AI and the impact it might have on their craft, and their industry.
But dive into design forums and communities, and you find a very different discussion around AI-powered upscaling. Designers seem energised and excited about the creative potential these tools open up.
AI-powered technology unlocks new creative and performance possibilities.
It removes image quality as a barrier, and allows designers to focus on what they do best – create visual lay-outs that make the biggest impact imaginable.
Ready to upgrade your set-up? Explore the latest NVIDIA GeForce RTX graphics cards and find the right fit for your rig – shop NVIDIA’s 50-series range here.