facial-recognition-ip-patents

How We Can See the Future of Facial Recognition

Artificial Intelligence Tags: Facial Recognition, InnovationQ, IP.com, natural language processing, Patent Search

The Questions

Facial recognition technology has generated recent buzz thanks to fun tools like FaceApp[1]. They can alter our present perception of ourselves and seemingly predict our future.

FaceApp and other facial recognition applications use artificial intelligence to get to “know” us. We snap our faces and upload them to databases where they wait to be matched for tagging in social media, opening mobile devices, even making digital transactions. It’s fun and handy. Facial recognition is also used by law enforcement agencies, from small cities to international airports, to track individuals’ movements and catch lawbreakers. It’s for our protection. In China, the government uses facial recognition technology to a much broader extent[2]. It is ubiquitous and purpose-driven.

As inventors forge ahead to uncover what else facial recognition does, where will the technology take us next?

How to Find the Answers

To investigate, we can search the intellectual property that is currently in play. InnovationQ helps answer questions like: What technologies do the most recent patents and applications represent? Which inventive paths are being plotted? Who is putting money into it?

And, rather than struggle with finding the exact keywords, we’ll use natural language, semantic-based searching to begin our quest. Similar to what facial recognition does for our image, InnovationQ’s semantic analysis does for our language. It analyzes real-life input, applies AI to identify meaning within the context, and then returns the best matches.

Let’s use the simple explanation of “facial recognition” that Wikipedia offers to create an InnovationQ search query:

A facial recognition system is a technology capable of identifying or verifying a person from a digital image or a video frame from a video source. There are multiple methods in which facial recognition systems work, but in general, they work by comparing selected facial features from a given image with faces within a database. It is also described as a Biometric Artificial Intelligence based application that can uniquely identify a person by analyzing patterns based on the person’s facial textures and shape.

Because this is a high-level description, we expect broader results. To keep the results recent, we can add a publication date filter of five years for the worldwide patents and patent applications.

Within a few seconds, InnovationQ produces a ranked list of over 11 million patents and applications. To answer our questions, we first look at the companies doing the most work in this area. Following a quick click, the system generates a graph of the assignees that shows how many documents they have in the results set.

Figure 1: Of the 1,500 most relevant results, these are the 15 assignees with the most patents and/or applications published in the last five years

Digimarc, Google, and Intel have the highest number of relevant documents. This is an interesting finding because each of these companies has a different core business, yet all are working on facial recognition.

Another simple click in InnovationQ presents a different perspective. Now, we see that EyeVerify has the most highly-relevant patents or applications. This could get us closer to the information we are looking for.

Figure 2: EyeVerify has the greatest number of patents and/or applications relevant to the query

What is EyeVerify working on? By rerunning the semantic query and just adding “EyeVerify” as a Boolean filter, we get a concentrated result set. Reviewing the top 20 terms from those documents gives us some insight as to what is happening. The larger, bolder terms most frequently appear in the results. The black font shows our query terms. The blue font shows non-query but prevalent terms.

Figure 3: The most prevalent terms in EyeVerify’s found documents

Based on a broad, simple query (that we copied from Wikipedia), we took just a few minutes and a couple of short steps to generate significant clues about the future of facial recognition. We’ve got some answers, but now we have more questions.

How are today’s innovators combining biometric patterns and facial details to feed AI? Here, EyeVerify applies ocular details. What’s next? Voice? Heart rate? Cholesterol levels? DNA? Is this a new path or a progression of an existing method?

What are the applications for this combination?

What power would it provide?

How to Find Out More

Run more pointed queries in InnovationQ. See the work that already exists. Use the visualizations and take advantage of the analytics. Keep searching, keep digging. New facial recognition technology is just one of the thousands of discoveries you can make through InnovationQ.

Whether you are an inventor, a researcher, a developer, or a patent attorney, IP.com’s semantic-based search and analytics tools can lead you to information recognition.

Contact IP.com to find out more.

[1] https://www.faceapp.com/?next=faceapp

[2] https://fortune.com/2018/10/28/in-china-facial-recognition-tech-is-watching-you/