Brand Detection: Event Sponsorship Analysis with querifai.ai

Discover the impact of event sponsorships with brand detection on our AI SaaS platform and identify brand logos in social media posts!

In this guide, you will learn how brand detection on querifai.ai can make your event sponsorship analysis on social media easy. Simply run the Brand Detection APIs 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 logo has appeared in those images. It automates the manual process of counting the logos from many images, thus saving the sponsorship manager time, effort, and chances of error.

Event Sponsorship Analysis

Event Sponsorship means advertising your brand by supporting an event financially in exchange for brand exposure. Events are excellent opportunities to connect your brand with the target audience. However, just sponsoring an event is not the end - after sponsoring your brand, you need to measure the impact of your sponsorship. That’s where you analyze how many times the event advertised your brand and what the impact was. One of the crucial parts of event sponsorship is brand detection, which involves identifying and tracking a brand's logo presence within an event. You can analyze your brand's visibility against competitors, assisting companies to benchmark their performance. Manually assessing the visibility of the company's logo in event photos on its social media page is time-consuming and may not provide a comprehensive analysis of the sponsorship's reach.

How querifai.ai Brand Detection Makes Your Event Sponsorship Analysis Easy on Social Media

Navigating the vast landscape of social media during event sponsorships can be overwhelming, but the querifai.ai Brand Detection feature makes all this easier for you. 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 scraped from social media posts. Although many existing online tools are available for brand detection, our platform stands out! querifai.ai provides:

A user-friendly no-code interface to analyze your images and to compare different brand-detection algorithms.
An API to integrate any feature of our no-code platform into your own working environment.

The sponsorship manager can scrap the images from the event’s social media page, upload them to our platform, and let our platform process them by identifying how many times each logo has appeared in their posts. Thus, our platform offers numerous advantages to the sponsorship manager.

Hard facts extractable from social media buzz
Time savings through automation
Reliable results from automated four-eye principle through parallel access to multiple AI services

Step-by-Step Guide to Run Brand Detection Through querifai.ai on Social Media Posts

Let’s go through a step-by-step guide demonstrating how to utilize our platform's brand detection feature. Anyone can do it!

Below are a few sample images of a famous brand like Google, sponsoring a conference event. As a sponsorship manager, you can visit the event's social media page, and download the images into a folder (For this example we have used the press-kits of Google and Microsoft). Now, the dataset is ready. So the next steps are:

Create an account on querifai.ai if you haven’t done so already:
Go to https://querifai.ai/home
Select Vision from the Use Cases panel.
From the Vision use cases, select Brand Detection.
Here, you have to upload the images from social media. Click on the Upload new dataset.
Give a title to your dataset and upload the files by clicking the Choose Files. Once uploaded, click on Create.
You will be navigated back to the existing datasets page. Select the newly uploaded dataset from the list and make sure to click all the services, i.e., Google, Clarifai, and Azure, then click Evaluate.
You can see the summary of the Use Case. Click on Proceed.
Here you go! In just a short amount of time, the brand names will be highlighted, and all the API’s will show the accuracy of the results.

Analyzing Multiple Cases on the API Results

Case 1: A simple situation, the logo is presented well-illuminated and isolated on a wall: All three APIs recognize the logo correctly.

Case 2: The logo is shown on a screen, the texture is thus noisy: Here, only Google was able to correctly identify the logo

Case 3: The logo is presented as a gem hanging outside on a sculpture. Curiously, Google does not recognize their own logo correctly. However, none of the APIs could match the brand correctly in this noisy setting.

Case 4: A skateboard with a light print of the Google logo. Clarifai recognizes this correctly.

Case 5: Interestingly, both Clarifai and Azure correctly identified Volkswagen’s logo with 83% and 68%, respectively: Results are more reliable with the consensus of multiple API’s results.

Case 6: All the API’s correctly identified Microsoft’s logo again, however, again, Google stands out by achieving an accuracy of 97%.

Case 7: Clarifai correctly detected the Google logo with an accuracy of 63%. While the other two APIs were not able to detect the logo.

Case 8: None of the APIs was able to detect any logo at all. It also provides a key finding that whenever the logos are small often go unnoticed by the AI. However, the issue can be resolved by applying preprocessing techniques to the input image, like zooming in.

Case 9: This time, the API was provided with the zoomed-in version of the same previous images. Now two of the APIs (Clarifai and Azure) correctly identified the Apple logo with an accuracy of 79% and 74% respectively.

Case 10: The Azure API has correctly identified Microsoft’s logo on the laptop, However, the other two APIs provided no results at all.

Conclusions from the Findings of Brand Detection Through Multiple APIs

Let’s discuss some interesting findings that you can observe from the results. You have to consider a few things after getting the results of brand detection.

The Random Nature of Bias Among Services

It's challenging to declare a clear winner in accuracy among the services. Surprisingly, Google's own logo was more accurately recognized by Clarifai than Google’s service in certain cases, like on a skateboard. This lack of a consistent bias suggests that each model has its strengths and weaknesses, making it sometimes suitable and other times not. Such randomness in performance indicates the need for a broad perspective when choosing a service for brand detection.

The Power of Combined API Results

When two or more APIs identify the same logo, the likelihood of accurate detection increases. For example, the Apple and Volkswagen logos were correctly identified by multiple services. This consensus is a robust indicator of the presence of a specific brand, providing a higher degree of confidence in the results.

Challenges with Small Logos

A significant limitation observed is the difficulty in detecting small logos. If the logos that are too small often go unnoticed by the AI. However, this issue can be resolved using various pre-processing techniques, such as image slicing and zooming, which will significantly enhance the detection rates.

Use Four-Eye Principle in Brand Detection

When it comes to identifying brands in images, it's important to remember that different APIs can give different results. Some do better with certain images, while others might be better with different ones. This is why using the "four eyes principle" in brand detection is quite important. This means you should always check your results with more than one source. It's best to combine the strengths of two or more APIs and blend AI's precision with human verification.

Obtain the Count of Each Logo After Brand Detection

After visually inspecting the detection rates and choosing one or more services for a specific set of images, the counting can be set up fully automatically in any environment. For this, querifai offers an easy-to-access REST API. The API will provide structured access to the result, including the count, unique labels, and more.

To make a call to the API, an API access token is needed. That token can be found in the Account Security section on the User’s page. Copy the token, as you will need it later in your code.

API Access token in account settings

You can call the API in any language. Below is the code for calling our API for the brand detection feature in Python.

Copy and paste the code, and you only need to fill in two things:
The directory containing the images: Replace the FOLDER_CONTAINING_THE_IMAGES with the actual directory.
API Access Token: Replace the YOUR_API_ACCESS_TOKEN with the token you copied from your account.
Run the code, and you will get the count of all the logos that appeared in those social media images.

Results from the Google API

Conclusion

In the dynamic field of event sponsorships, querifai.ai Brand Detection poses as a practical and time-efficient solution for sponsorship managers without complicating it. querifai.ai doesn't just make your job easier - it transforms how you approach event analysis. It's not just about data; it's about making informed decisions and maximizing your brand's impact on your sponsorships. So, are you ready to revolutionize your event sponsorships? Utilize the querifai.ai Brand Detection and watch your brand's impact. Take advantage of the power of data-driven decisions and stress-free analysis. Your brand's success is just a click away. Join querifai now and take your event sponsorships to the next level!