AI special May 12, 2024
Derek Milne, a commercial pictometrist at Pixometry, says AI is significantly streamlining workflows and improving images within editorial and photography teams.
Q: What did you learn about using AI in image creation and editing?
A: An AI imaging engine integrated with automatic image optimization workflows provides publishers with a powerful and incredibly efficient toolset. There is little need for a timely onboarding or training process for users, resulting in almost immediate benefits in terms of quality and efficiency.
We’ve heard a lot about Adobe Firefly and its features, but images still require a separate manual process. Prompt-based image generation is great for one-off images, but the results are still questionable. That said, Firefly’s ability to extend image backgrounds to fill space is probably the most valuable tool for the image department.
However, with a focus on automation and AI batch processing, there are even more possibilities that can be realized beyond the realm of Adobe. Automation that does much of the heavy lifting for image optimization requirements provides core functionality and superior results, freeing up time for the image department to focus on the creative side.
Q: In which use cases has AI proven to be most effective?
A: Major UK newspaper and magazine publishers have successfully leveraged AI imaging technology to streamline their imaging processes. We strived to improve efficiency and quality across all channels by producing daily, weekly, and monthly publications. The phased implementation established a software platform for leveraging AI tools. Some AI tools are for specific departments, others for all titles, maximizing the potential of the combined technologies.
The foundation of the imaging workflow was image enhancement software that efficiently processed large numbers of images while meeting the art director’s exacting quality standards.
The platform adapts to each title’s style, creates renditions for different printing processes, and digital I needed to generate a version.
Image automation quickly achieved significant efficiency gains by saving time in image processing and layout, standardizing image quality, and improving print quality. The imaging team now has more time to focus on the images that matter, further improving the overall quality of the product.
The quality of online images has improved significantly, making them more attractive. Content creators who typically lack the tools and skills to enhance images can now create powerful images that attract readers.
Once the automation platform was established, we began integrating it with an AI batch processing engine for background removal and image recognition.
AI-based automatic background removal has been available for over five years. Incredibly accurate and fast, this technology is ideal for high-volume workflows.
AI services are integrated into publisher workflows and managed by an extensible platform. Users send images for cutouts to the cloud, and files are returned within seconds, ready to be placed. No Photoshop skills required.
Over 4,000 images are processed for cutouts each month, with a success rate of approximately 95%. Success rate means that the image is ready to be placed without additional editing. These go beyond simple headshots to include sports, products, animals, vehicles, and more, with intricate details like the strings on a tennis racket perfectly cropped.
Image quality and time savings have led publishers to use more cutouts in all forms of content, producing more creative and attention-grabbing images.
The third element of AI technology is the evolving science of image understanding, which aims to identify and tag image content.
We all know the photo of a child eating ice cream on Westminster Bridge with Big Ben in the background, but on a computer it’s just a pixel. Image understanding adds appropriate keywords to the file’s metadata, such as “children,” “bridge,” and “ice cream,” and can also add GPS coordinates to identify Big Ben.
Deployed as an integral part of the image processing workflow, the usage and possibilities of this technology are blurred across different roles and departments. During the image enhancement process, the generated keywords are utilized by the software to fine-tune the correction criteria, resulting in a more refined and powerful image. Additionally, tags are added to filenames to improve image SEO for your online team.
Across the publisher ecosystem, these added keywords provide significant discoverability benefits within content management systems, making your images more reusable.
Further business opportunities for the archive are being discussed with the digital content team. While the digital team aims to improve loading speed, image SEO, and discoverability, the archive focuses on monetizing its large image collection with discoverability as a key factor.
3 important tips for best practices
Take advantage of image automation to handle time-consuming and repetitive tasks. Step by step. You don’t have to do everything at once. Which toolset has the most benefit right now? Look at your space. Technology is constantly evolving with new tools.
Derek and other AI Special contributors will be joining us for the “AI Special – Q&A” webinar on Tuesday, January 28th. Click here for more information and to register.
For more than 25 years, Pixometry’s advanced image enhancement software has enhanced the image workflows of publishers around the world. Our continuously evolving software incorporates the latest AI imaging technology, making it perfect for enhancing and enriching images to engage readers in print and digital media.
Email: derek.milne@pixometry.com
Website: www.pixometry.com
This article was published in the AI feature published by InPublishing in December 2024. Click here to see other articles in this feature.
