Logo Identifier: Build A Logo Detector in 2 Minutes
Your logo is a valuable brand asset. But monitoring its usage across the web can be near impossible. What if people are using it incorrectly, or your ad designer placed it upside-down?
The good news is, real-time logo identification (also called logo recognition or detection) allows you to automatically double-check that a logo is being used correctly. Use cases include:
- Your ads correctly display your brand’s logo and orientation
- Someone isn’t editing or defacing your logo
- Someone isn’t using your logo on counterfeit goods
In this article, we’ll show you how to create a logo identifier in just minutes.
How does logo identification work?
Logo identification uses machine learning image classification to recognize the unique visual characteristics of your company’s logo. After it’s trained, the model will be able to flag instances where a logo doesn’t exist or is being used incorrectly.
Training these classifiers can take time, as you need a number of logo variations to train the model. The good news is, we’ve built a tool that allows you to launch a logo identifier with just a single image.
See it in action
Upload a food-related logo below (such as Dominos, Wheat Thins, Olive Garden, etc.) and see if it’s identified correctly. While the below is trained with logos from large food brands, any brand can build a logo identifier for their own logo.
How to build a logo identifier in 5 steps
1. Find one image logo file
Ideally this file would:
- Be in .png format
- Have transparent background
- Be at least 300px wide
- Be a color logo
2. Upload that file Below
After you upload it, our system will create a folder with a variety of small adjustments to train the image classifier.
3. Set up your Nyckel account (it’s free)
- Sign-up for free
- Create an “Image Classification” function named “Logo Identifier”
4. Set up the classifier
- Create a new Image Classification function
- Call it “Logo Identifier”
- Add three labels: “Good Logo”, “Bad Logo”, “No Logo”
- Go to the Import tab
- Upload the batches in three sets (Good, Bad, No)
- For each set, drag the images to the upload box, then tag with the right logo
It should take about 10-20s to train. Go to the Test tab and insert an image to see how it’s classified.
There you go - in less than 2 minutes you have your own logo classifier set-up!
There are many ways to integrate this logo classifier into your wokflow, including:
- Before your team uploads banner ads in bulk to an ad platform, the images get sent to the classifier. Any that are flagged get manual review (via our API)
- You could monitor social media usage of your brand logo by pulling user-generated images and sending them to Nyckel (via the API or Zapier)
- If your assets are stored in a Digital Asset Manager (DAM) or a similar tool, you could have those images be sent to Nyckel for validation upon upload (via the API or Zapier)
- You could give designers a web portal where they can double-check if their image passes brand standards (via an embedded JS widget)
Iterate and Improve
Nyckel provides a feature called Invoke Capture that will automatically flag images for manual review. Spending just 5-10 minutes a week manually tagging these will improve the model over time.
Nyckel provides the flexibility, accuracy, and adaptability that businesses need to create a reliable logo identifier. If you’d like to take the next step to safeguarding your brand, sign up for a free Nyckel account to create your custom logo identifier. Run into any issues along the way? Reach out to us at any time.