Pongr has announced the launch of ImagePulse, a visual search engine that measures brand sentiment by indexing and ranking the endless stream of brand photos taken by consumers on their mobiles and uploaded to social networking and other websites.

Pongr notes that many people who now share photos of their favourite foods, fashions and products on the web are virtually invisible. Consumers snap and share billions of pictures every day with their mobile phones. As the volume of these photos continues to rise, the quality and frequency of tagging is going down. Informal brand ambassadors aren’t aiming to advertise – their impulses to share brand-themed photos, which are social, and thus, do not come labelled with hashtags.

Pongr’s image-recognition platform “sees” company logos regardless if they are tagged or not. Additionally, each photo is given a Purchase Intent Score based on an additional layer of text analysis for buying words. Companies can also learn what else excites their most loyal customers with “conjunctive interspend” data that tracks engagement with other favourite brands. For example, advertisers may find that a high percentage of adidas fans own iPods, drink Starbucks, watch MTV and shop at Target.

“A picture is much more indicative of product interest than a Tweet,” says Pongr CEO Jamie Thompson. “There are lots of sentiment analysis tools out there, but none of them offer visual sentiment analysis. There’s now a new way to meaningfully tap into existing customer behaviour, speed up collection of data, and provide a direct response.”

Voluntary product endorsements organically happen as a routine part of life, says Pongr. Examples of fan photos include iPad lovers photographing themselves in the Apple Store’ Hard Rock Café diners hamming it up in their logo t-shirts; or a fashion hound posing in front of a Versace or Gucci shop window.

“Advertisers can use ImagePulse  to learn how, when and where their unofficial brand ambassadors – and the competition’s – are marketing their products in social networks,” says Thompson. “We also can track down fan photos anywhere they are shared: on blogs, social media, mobile phone apps and online photo albums.”

In a preliminary search of Twitter photos (tagged and untagged) over a recent five-week period, ImagePulse collected and analyzed more than 80,000 fan photos across 47 sample brands. Pongr notes that this fan-generated content represents only a tiny fraction of the visual brand sentiment expressed across the web. The top 10 brand logos found on Twitter were:

1. Google: 22,483
2. Apple:14,452
3. Android: 11,127
4. Starbucks: 10,713
5. Coca-Cola: 6,339
6. Nike: 5,754
7. Adidas: 5,244
8. Nintendo: 5,094
9. McDonalds: 3,435
10. Converse: 3,215