Welcome to Visualize IP. We are former USPTO Examiners, patent searchers, patent attorneys, and data scientists. Our mission is to make high-end patent research accessible, affordable, and on-demand - all while matching or exceeding human performance (more on this later).

My patent career started as a USPTO Examiner. I can remember cases where I would be searching for a particular image, leafing through thousands of patent images. Manually searching images was grueling, and time-intensive. Years later, I founded SPSG, a Southern California-based patent search firm. At SPSG, we noticed a considerable uptick in design patent-related searches - from FTO and invalidity, to patentability. The value of design rights was increasing. Design patent litigation damages were rising and design protection was becoming the centerpiece of many patent portfolios.

We noticed that for any manual design search, our team would reach fatigue or budget limits after reviewing ~30,000 images. That’s a lot of images. We also know the accuracy of design searching is 1:1 with the number of images searched. So, the more images searched, the more accurate the search. This “more the merrier” aspect of design searching is unique. The most accurate design search is one that searches millions of designs.

We also noticed design searching was binary in nature - an image is similar, or it isn’t. There is no grey area. I had a working understanding of machine learning, and had already seen various image similarity tools. I realized we had 3 key ingredients:

Domain expertise in design search
Valuable training data
A key partner in SPSG

I took the leap in 2019, and Visualize IP (“VIP”) was born. We are a patent SaaS company, pioneering computer vision patent search.

Like SPSG, VIP made quality a cornerstone. VIP would need to, at a minimum, match SPSG’s manual searching, and ideally surpass SPSG. In 2019, there were several patent SaaS companies who offered image similarity search. The results were underwhelming. We could tell they were utilizing generic, off-the-shelf image similarity approaches.

We wanted to be the first to get it right. This would become the ethos of VIP. We would spend almost 2 years in R&D. Relentlessly testing and improving. We eventually created an ensemble of algorithms that would succeed at every benchmark, consistently matching and/or beating SPSG manual searching. This moment is when we knew we developed a truly disruptive technology. Design searching would be accessible and affordable to everyone, and would be of the highest quality. We are only just beginning, thank you for coming along for the ride.

-Eugene Lhymn, CEO and Founder

Getting it Right, the First Time.