In order to defend your perimeter, you must first be able to define it. To do this, OnDOMAIN enables you to easily upload all variants of your organisation’s brand assets to its logo management library. It will then scan your logos and train itself to recognise them across the web.
Once you have uploaded your logos, OnDOMAIN uses computer vision and machine learning to scan the web for both legitimate and illegitimate use of your organisation’s brand assets. When it determines that assets are being illegitimately used, these findings act as an additional dimension of classification to detect brand misuse or fraud and increase the speed with which a takedown can be initiated.
How do I use this feature?
It takes only a few clicks to upload your logo(s):
Click on the + Add logo button.
Select a logo image file and provide a class name, either from the existing list or a completely different one.
Click on the ADD LOGO button to confirm the submission.
Please note:
To help train our AI to detect your logos, define your own classes to group similar logos together. To learn more about classes, scroll down to read more details.
Please note that it will take up to 48 hours for our AI to scan and process your logo(s).
Logo images must have a minimum width or height of 128px for the best quality from our logo detection service: Supported image formats are SVG, PNG, and JPEG.
Please consider cropping any void background from the image file before submitting the logo, for example:
Tip to find an SVG log:
For large organisations with a wikipedia page, there is almost always an SVG logo in the top right corner under the company name.
The homepage sometimes has SVG logo as well. For example https://redsift.com/, save the page, look inside the downloaded folder, the logos are in that list of files.
What are classes?
A class is a group of logos related to the same brand name, i.e. both the brand name and its visual logos are different from other brands.
For instance: 5 logos - OnINBOX,OnDOMAIN,OnDMARC and Red Sift
The first three logos need to be separated in their own individual Class as they represent different products.
The two remaining logos (RedSift) stay under one single Class because they’re semantically about a single brand. (Note the difference under Logo-REDSIFT; Two logos were uploaded under the same Class).
Do not upload identical logos of different colours - AI is capable to detect those.
Ex:
What happens if OnDOMAIN detects illegitimate use of my organisation’s logos?
OnDOMAIN will warn you of the presence of a detected logo by flagging a red warning triangle in the "Logo annotations present" column.
The discovery of illegitimately used logos will also be used to determine the risk rating of a lookalike domain.
To learn more about risk ratings OnDOMAIN, please follow the link:
Please refer to the article below for the next step: