Q: What have you learnt about using AI in image creation & editing?
A: AI imaging engines integrated with automated image optimisation workflows offer a powerful and incredibly efficient toolset to the publisher. There’s little need for timely onboarding and training processes for the users, and the results yield almost instant benefits in terms of quality and efficiency.
Whilst we hear a lot about Adobe Firefly and its capabilities, it remains a manual and individual process for images. Prompt-based image generation can be excellent for one-off images, yet there are still questions over results. That said, Firefly’s ability to extend image backgrounds to fill space is arguably its most valuable tool for imaging departments.
However, there are more opportunities to be realised beyond Adobe’s realm by focusing on automation and AI batch processing. Performing much of the heavy lifting of image optimisation requirements, automation provides core functionality and excellent results, freeing up the imaging department’s time to focus on the creative side of their craft.
Q: In which use-case has AI proved most effective?
A: A leading UK newspaper and magazine publisher successfully utilises AI imaging technology to streamline their imaging process. Producing daily, weekly and monthly publications, they looked to boost efficiency and quality across all channels. A phased implementation established a software platform to leverage AI tools, some for specific departments and others for all titles, maximising the potential of combined technologies.
The cornerstone of the imaging workflow was image enhancement software that met the rigorous quality standards of art directors while efficiently processing large volumes of images.
The platform needed to handle a wide variety of images, from professional to social media shots and everything else, while adapting to each title’s style, creating renditions for different print processes and generating digital versions.
Image automation quickly achieved major efficiencies, saving time in imaging and layout, standardising image quality and enhancing print quality. Imaging teams gained more time to focus on critical images, further raising the product’s overall quality.
Online images saw significant improvements in quality, making them more engaging. Content creators, typically lacking the tools or skills to enhance images, can now produce powerful images that attract readers.
With the automation platform established, integration with AI batch processing engines for background removal and image recognition was initiated.
AI based automated background removal has been available for over five years. Incredibly precise and fast, the technology is perfectly suited for high volume workflows.
Integrated into the publisher’s workflow, the AI service is managed by the enhancement platform. Users send images for cut-out to the cloud, and within seconds, files return ready for placement — no Photoshop skills required.
Upwards of 4,000 images a month are processed for cut-outs, with a success rate of ~95%. By success rate, I mean images are ready for placement with no further edits. These aren’t just simple headshots, they include sports, products, animals, vehicles, and more, with even complex details like tennis racquet strings cut out perfectly.
The image quality and time savings encouraged the publisher to use more cut-outs across all forms of content, generating more creative, attention-grabbing imagery.
The third component of AI technology is the evolving science of image understanding, which aims to identify and tag image content.
While we all recognise a photo of a child with an ice cream on Westminster Bridge with Big Ben in the background, to a computer, it’s just pixels. Image understanding adds appropriate keywords like ‘child’, ‘bridge’, ‘ice cream,’ to the file’s metadata and can even identify Big Ben, adding GPS coordinates.
Deployed as an integral part of an imaging workflow, the use and potential of this technology is blurred across different roles and departments. For the image enhancement process, the generated keywords are utilised by the software to fine tune enhancement criteria, resulting in more finessed and powerful looking images. Additionally, the tags are added to the filename for improving image SEO for the online teams.
Across the publisher’s ecosystem, these added keywords give a huge benefit for discoverability within the content management system enabling images to be utilised again and again.
Additional business opportunities for digital content teams and archive are in discussion. The digital teams aim to improve load speeds, image SEO and discoverability, while the archive focuses on monetising their vast image collection with discoverability as a key factor.
Three best practice top tips
- Embrace image automation for the heavy lifting of time-consuming repetitive tasks.
- Step by step. It doesn’t need to be everything all at once. Which toolset offers the most benefit now?
- Watch the space. The technology is constantly evolving with new tools.
Derek and the other contributors to our AI Special will take part in an ‘AI Special – Q&A’ webinar on Tuesday, 28 January. Click here for more information and to register.
For over 25 years, Pixometry’s advanced image enhancement software has been powering the imaging workflows of publishers worldwide. Continuously evolving, our software now incorporates the latest AI imaging technologies to enhance and enrich images, perfect for engaging readers in print and digital media.
Email: derek.milne@pixometry.com
Website: www.pixometry.com
This article was included in the AI Special, published by InPublishing in December 2024. Click here to see the other articles in this special feature.
