Brand Detection: Event sponsorship Analysis with querifai

Explore the influence of event sponsorships through brand recognition on the AI-powered platform querifai.ai, and pinpoint brand logos within social media content.

Please note: This is a shorter version of the original article, the full version can be found here.

Event sponsorship

Event Sponsorship is more than just funding an event for brand exposure; it’s about creating a connection with your audience. However, the real challenge begins post-sponsorship: measuring the impact. A key aspect here is ‘brand detection’ — tracking your brand’s logo and comparing visibility with competitors. Traditional methods of assessing this on social media can be time-consuming and may not be practical.

Luckily, the querifai.ai Brand Detection feature makes all this easier. querifai.ai utilizes advanced artificial intelligence and machine learning algorithms by different vendors like Google, Clarifai, and Azure to identify and track brand presence on the images. Simply run the Brand Detection Use case on the event’s social media images. querifai.ai helps you to detect all your brand logos on those images and get the count of how many times your and the competitor’s logo has appeared in those images.

A sample experiment, where a selection of Google’s and Microsoft presskit images were given to the brand detection API yielded good results.

Selected Findings from the Brand Detection Experiment through querifai.ai

Case 1: All APIs correctly Identified the Logo

In a straightforward scenario, the logo is clearly displayed, well-lit, and set apart on a wall. Each of the three APIs successfully identify the logo.

Case 2: Single API yielded the correct result

The logo appears on a screen with a grainy texture, and in this instance, only Google’s API accurately recognizes the logo.

Case 3: Google Misclassified Its Own Logo

A skateboard features a faint print of the Google logo. In this case, Clarifai alone identifies it.

Case 4: Two out of Three APIs were Successful

Both Clarifai and Azure correctly identify the logo with 83% and 68%, respectively: confirmation of brand presence is more reliable with the consensus of multiple API’s results.

Case 5: Small Logo — No API provided Results

The example photo shows a scene from a coffee shop, several tables with MacBooks on them. However, none of the APIs detect the Apple logos (at the back of the screen) in this scene. This shows that logos that are small can escape AI detection. However, this challenge can be resolved by using preprocessing methods on the input images, such as zooming in, slicing, etc.

Insights from Brand Detection with Multiple APIs

Varied Accuracy Across Services

No single service consistently outperforms others in brand detection accuracy. Even, different services, like Clarifai, sometimes recognize brands like Google’s logo better than Google’s own service, depending on the context (e.g., on a skateboard). This inconsistency highlights the importance of considering various options, as each service has unique strengths.

Handling Small Logos

Small logos pose a challenge for AI detection, often remaining undetected. However, techniques such as image slicing and zooming can improve detection effectiveness for these smaller logos.

Implementing the Four-Eye Principle

Relying on a single API for brand detection can lead to varied results. Different APIs may perform well with specific images. It’s always useful to utilize the “four-eye principle” — verifying results through multiple sources to combine the precision of AI with the reliability of human oversight.

Complete Guide to Brand Detection on Our Platform with API

For a complete step-by-step guide to brand detection on our platform and the detailed results of using querifai.ai in Event Sponsorship Analysis, check out the full article.

The Python code for using our API is also provided with the complete guide, which gives much deeper insights like the unique labels, count of all the logos, etc. Discover how querifai.ai can transform your brand’s visibility analysis and benchmark performance against competitors.