ADP
Year: 2019
AI-Generated Branding
Agency: Havas New York
This project represents an early exploration of artificial intelligence through a client collaboration with ADP. I was tasked with examining how AI could be used to reimagine and evolve their visual identity.
Using a technique known as style transfer, I trained on a unique image to generate hundreds of AI-produced visuals, including both images and videos, as part of the brand's creative transformation.
Process
Style Transfer is technique of recomposing images in the style of another image. It takes two inputs—a content image and a style reference image—and blends them using convolutional neural networks.
Using style transfer, we developed the Uplifting style as our style reference image. It incorporates tones and organic shapes drawn from ADP's new visual identity and is available in three color schemes: Purple, Rose, and Tan. Training the model on each input took approximately 8–9 hours.
While the training process required several hours, applying the style to an image took only a couple of seconds. This efficiency was crucial, as we needed scalable tools capable of generating content at the volume required for a project of this scale.
Videos
Our Style Transfer technique is built upon Manuel Ruder’s fast-artistic-videos method. It offered an implementation of video style transfer with improved frame-to-frame stability by adding optical flow*. This resulted in generating consistent and stable stylized video sequences even in cases with large motion and strong occlusion.
Stylizing videos was a complex process due to optical flow computation. Training one input took 16 hrs and stylizing a 30 second video took approx. 4 hours
*Optical flow refers to the visible motion of an object and the apparent 'flow' of pixels in an image.
Brand Launch
Uplifting fits perfectly within the overall design guidelines for ADP. It is never used by itself. Stylized images and videos can be composed of a mixture between our shapes, our icons library, and our custom typography - giving it overall a polished look.
Early Experiments
Before Uplifting, we experimented with many different models and styles. One technique we tested was built upon Hang Zhang’s Multi-style Generative Network for Real-time Transfer method.
We tested very fine patterns composed of lines and grids with variations in size and thickness.
This next test was trained on a set of early ADP icons. They are composed of organic lines and separated blocks of colour. It replicates a more minimalist design approach and a broad use of white space.
Continuing along the previous path, we tested organic figures and sets of colors, but this time used larger/bolder shapes and got rid of our strokes.
Uplifting Style (Early Version): We found that the variations in tones and overlap of organic and delicate geometric shapes yielded the most beautiful results.These were our very first passes with Uplifting. This style continued to improve and morph into the final version that we see today.
2019 – SXSW