Blog

FraudWatch International Image Recognition Technology

FraudWatch International Combats Brand Abuse with Image Recognition

posted by: FraudWatch International date: Apr 12, 2019 category: All, Brand Abuse, Brand Protection comments: 0

FraudWatch International has made significant headway in its R&D labs with real-time, high-velocity image recognition scanning. So far it has yielded encouraging results around accuracy of brand detection from a huge volume of data sources and across multiple brands.

Image Recognition

Simply stated, image recognition is the ability to detect an image. Basically, image recognition technology works by getting a computer to see and recognize certain images. As a part of Computer Vision, image recognition involves programming computers to recognize digital images or videos. Although this may sound like a simple task, teaching a machine to mimic the human visual system is extremely complex. 

By having access to an API, the software searches through data for client’s logos, and raises an alert if it detects something of interest or out of the ordinary. This technology recognizes clients’ brands and logos, increases the detection rate, and reduces manual work. When it comes to brand protection, image recognition technology vastly improves the process of detecting counterfeit logos and images, which seek to impersonate a legitimate brand to defraud people. 

As well as recognizing images, this technology also facilitates machine learning. While the program analyzes images, it learns and becomes more efficient at doing its job. Through this machine learning process, the technology can also set rules and also prioritize brand infringements.

One of the greatest benefits of this technology, however, is that it works around-the-clock, scanning websites for logo abuse and brand infringement. In the past, detecting counterfeit logos and misuse of a client’s brand, was a labor-intensive task for human eyes. In contrast, a machine that is capable of detecting images can do this work much faster than a team of humans ever could. 

Working Faster for Better Service

Image recognition will become a key component to FraudWatch International’s brand protection service, and will help us to provide better brand protection for each client. Our Security Operations Center (SOC) receives large amounts of data on a daily basis, and manually processing it all is extremely time-consuming.

By utilizing image recognition, FraudWatch International will be able to improve image detection rates. On top of that, the SOC will be able to detect malicious content in a more-timely manner. This technology will also detect images on sub-pages, in addition to the base domain, therefore extending FraudWatch’s ability to detect brand abuse.

To date, FraudWatch International has produced successful results with image recognition tests and yielded positive results on 20 brands. In fact, incidents are being logged on a regular basis. As a result, this has improved FraudWatch’s detection rate for its clients, which will further enhance FraudWatch International’s industry-leading takedown times.

Safeguarding Brands and Reputations

Building a reputable, trustworthy brand takes years. Yet, a single serious case of brand abuse can severely damage a brand’s reputation. There are several ways that fraudsters perpetrate Online Brand Abuse, such as: brand impersonation, domain name abuse, and illegitimately affiliating themselves with trusted brands. As a result, online brand abuse can lead to loss of revenue, a damaged reputation, and aggravated customers. 

What’s more, misuse of a brand infringes on trademarks by failing to respect copyrighted content. Although sharing a brand’s original content might not be malicious, it is still an infringement. Image recognition helps prevent unauthorized or improper use of brand’s logos.

One of FraudWatch International’s major focuses is online brand protection. In addition to the manual work already being done by our teams, harnessing image recognition extends our ability to protect its clients’ brands, along with their customers.

This service is something FraudWatch International plans to roll out to better serve all of its clients. However, extending this service to all clients won’t be immediate. Due to the number of clients and the variety of logos (different sizes, colors, shapes and places), collecting logos, labeling them, and training the image recognition model will take time.

As product development continues, we expect it to have a soft launch into our production system and SOC’s work flow. Our team will perform human verification of the results that this image recognition technology produces. Once we are satisfied that the level of performance and accuracy meets our standards in providing excellent service, we will release it into production, alongside our other detection systems, to bolster our detection methods. To learn more about FraudWatch International’s image recognition initiative, contact us.

Comments are closed.