The New World of Computer Vision

By Sara Musselman

Double Vision

Imagine watching the NBA Finals – Game 5. LeBron James is on the court driving to score mid-air with seconds left. He shoots and is fouled. Moments later, millions of eyes are on LeBron's face as he stands on the line with a chance to score for the Lakers and win the game. Time slows to the pace of anticipation. His focus is our focus, his vision is our vision.

Now, compare our vision to computer vision. What would a computer see in this game, frame by frame? To start, computer vision today does not see through the lens of loyalty to a team or hold any emotional hope for the win. It does collect data in every moment objectively.

Simply put, computer vision has the capacity to decipher millions of elements the human eye would miss, from the movements of each fan in the crowd to the brands and sponsors in the arena, as well as the sentiment of every player's face – simultaneously.

Jason Mills, executive director and head of customer engineering at Google Cloud Industries, uses this metaphor to explain computer vision and goes on to state, "This information is very powerful for people in the video industry." I completely agree.

Our Sight Extended

In a well-known experiment, researchers observed subjects watching a video of people playing basketball. They asked the subjects to count the number of passes between players. When the video was over, the subjects were asked if they noticed anything unusual happening in the video. More than half of the participants noticed nothing strange, not even the person dressed in a gorilla suit clearly marching across the court.

This study demonstrates "inattentional blindness," a human trait that prioritizes focus so much that we miss odd and obvious things in plain sight. Inattentional blindness helps us understand context and follow a story, but it also causes us to be blind to what is in front of us as well.

Inattentional blindness affects decision-making in the creative industry, influencing the creative we make for others to see. For those of us in the creative advertising and marketing fields, inattentional blindness limited our success. Until now.

Clarity Through Storytelling

Computer vision will revolutionize the creative process. Creative professionals have long held the unenviable position of reading the tea leaves and making a decision about something that was inherently unmeasurable.

For example, consider testing two different ad designs for a sneaker brand. From the top-line performance, we can only see that one ad did better than the other, but are not sure why. Typically, it had been up to the creative team to divine the reason, using professional opinions and experience to determine why one might be more appealing than the other.

Computer vision uncovers a trove of insights. It might reveal that one campaign performed better due to more saturated colors, bolder text, and an actor's gaze. Like the dial on a safe, this combination of information unlocks potential new mandatories, guardrails, and guidance for optimization.

Small details make a big difference in how creative stories are received. Much more informed decisions can be made about how to design creative or change media plans to favor better performance and connection with audiences.

Computer vision will elevate and empower brands. Brand standards are based on a variety of requirements from color and logo use to product placement, content, and copy. How those brand standards affect creative performance has never been measurable in any level of detail.

With computer vision, brand standards can be tested on a granular level, and immediate insights using creative data will tell us if they are positively or negatively impacting the brand's goals. With data and insights, the creative team gains a level of objectivity that can help brands move beyond assumptions, fears, and long-held beliefs.

Seeing Is Believing

It's always harder to argue when data is involved. Computer vision provides certainty that we have historically lacked. Rather than restricting the creative process, computer vision enhances it. We in the creative industry can spend less time defending our ideas and more time innovating with confidence.

Art and science are constantly in motion. With computer vision, it's possible to monitor those changes at a granular level. Regional tastes, demographic differences in opinion, cultural differences, all these things can be connected to performance. Designs can be altered to be more appealing and altered again as time goes on.

Audiences will always naturally have personal associations and biases, but by leveraging technologies like computer vision, we can connect with audiences in ways we never could have using our own vision alone.

MIT Professor Sinan Aral calls this new data set a superpower for creators. With creative data derived from computer vision, we can now augment our storytelling for the better, based on broader computer vision-driven insights, and connect with audiences in more creative and relevant ways. We eliminate the elements we thought could matter but really don't and build upon opportunities that we may have missed. Our human sight is extended through the application of computer vision. Beyond our blind spots, we will see differently.

Life always comes with a backstory, and creatives are adept at telling that story. Knowing how to frame that story as effectively as possible for every audience takes computer vision.

The views and opinions expressed are solely those of the contributor and do not necessarily reflect the official position of the ANA or imply endorsement from the ANA.

Sara Musselman is the executive creative director at VidMob.