Detecting Iconic Brand Assets

Why distinguish between iconic and text-based brand assets?

For junbi to calculate your Brand Attention score, it's essential that we can detect all instances of your brand, whether it's a text-based or iconic-based brand asset. While you and I might acknowledge iconic and text-based elements of a brand as the same, computers see them a little differently, and we need to account for this.

Detecting an iconic brand asset is a little bit trickier, so to do this, we have employed a unique method to detect iconic brand assets using what's known as segmentation. This is a computer vision term which means that we're able to get a much more accurate prediction about where the brand is located within an ad.

Take a look at the example below to see how it works:

ICONIC-BASED BRAND DETECTION

In the above example, you can see the green shape surrounding the logo which pops in at different moments of the ad. Compared to text-based logo detection, the result is a little different, as you notice with text detection, the brand detection is performed by drawing a box around the brand:

TEXT-BASED BRAND DETECTION

How to use iconic brand detection in junbi

Training an iconic model isn't easy, and as we want to incorporate elements specific to your brand we always train a completely tailored and custom model. If a custom model is included in your subscription we will work together with you and the rest of the team to make the model available to you as soon as possible. This is covered as part of our customer success program.

To use iconic brand detection, simply upload your ad, then in the next step when entering your ad details, select the brand you wish to test from the drop-down menu.

Select the option with the sub-heading 'Iconic and text based brands'. This will ensure that our iconic logo detection model is run on your ad. If you wish to simply use text detection (which is faster and more accurate for text-based logos) simply select the text-based option from the drop down list.