Episode 12

Stewart in the Studio

Computer Vision – Revolutionizing the Appraisal Industry

Stewart in the Studio, hosted by Marvin Stone, SVP, Strategic Initiatives, is a fast-paced monthly podcast focused on issues vital to the mortgage lending industry. Tune in each month as Marvin and industry thought leaders discuss important trends and timely topics.

To speak to a Stewart Lender Services expert, please fill out the form to the right – we will get back to you shortly.

Stewart in the Studio Form

  • reCAPTCHA helps prevent automated form spam. The submit button will be disabled until you complete the CAPTCHA.
  • This field is for validation purposes and should be left unchanged.

How to Listen?

Episode Notes

  • Guest speaker, Matt Jenkins, VP Valuation Services for Stewart Valuation Intelligence
  • Computer vision is being used to detect all the features in a property and run analysis on the property and the comparables.
  • Learn how computer vision is being used by SVI to improve throughput and accuracy.

Speaker Bios

Speakers for this episode include:

Marvin Stone

As Senior Vice President, Director of Strategic Initiatives for Stewart Lender Services, Marvin Stone is working on the digital transformation of the transaction process for Stewart’s full range of services that cover the entire mortgage lifecycle. He also contributes to industry technology by participating in MBA’s MISMO Title and Settlement Community of Practice and being part of Stewart’s generative AI council.

Stone has been with Stewart since 2007 and has managed various technology, process and compliance initiatives for the direct, agency and centralized title divisions. Before joining Stewart, he was CIO for a large title company on the West Coast and held strategic roles with other national underwriters and mortgage lenders.

Stone is a frequent speaker at industry events and is sought out for his commentary on industry trends.

Episode 12 Transcript

View Full Transcript

E12: Computer Vision – Revolutionizing the Appraisal Industry

Marvin: Hey Matt, it’s great to meet with you today. So when we talk about computer vision, what exactly is that?

Matt: Computer vision is really image detection. And it’s helpful to utilize computer vision to detect all the features in a property and run analysis on the property and the comparables.

Marvin: When you’re talking about computer vision at Stewart Valuation Intelligence, has this been a real area of focus technology wise?

Matt: Yeah, so this is a key initiative for our technology group and our operation in general.

Matt: Really, what we’re focused on is trying to bring that computer vision forward so that we can make the right decisions and we can improve, on our throughput and our accuracy.

Marvin: Where in the Stewart Valuation Intelligence line do you actually use computer vision?

Marvin: Where does it fit in?

Matt: It’s a computer visions used operationally in all of our forms and our quality control process. We’ve got a range of valuation products from BPOs all the way through the full traditional appraisals and image detection is really critical in that to make sure that we’re comparing like properties and that we’re taking into consideration the condition of those properties and providing that analysis to our valuation providers, but also our quality control departments. So, operationally from quality control to form fulfillment, that’s where we see the benefit in computer vision and the analysis and the insights that it can provide.

Marvin: So, it’s used throughout the entire process, and then Stewart Valuation also has the VALIDITY app, and I’m assuming computer vision is incorporated in the VALIDITY app, as well.

Matt: Yes, VALIDITY Pro, which is our application for the property data inspections, or valuation modernization, was really the pioneer for the implementation of computer vision in our platform and our forms. That’s where we kicked it off.

Marvin: So the app actually is used for inspections and then those pictures that the inspector takes are all run through the SVI computer vision.

Matt: Exactly right. Yep. So the app is used for the inspection and as the inspector goes through the property, the computer vision can help determine exactly what is in the property the fixtures, the features, the amenities, the double sinks, the stoves, tile floors what types of carpet, what types of walls, stone counters.

Matt: It can also help you make sure that you view the property in terms of the exact inspection that we have to deliver. So floor to ceiling photos, making sure that you’ve accounted for everything that the computer vision sees in terms of outbuildings, various sides of the property. And it drives efficiency for the inspector and a lot of accuracy for the result of the product and our quality control group as well.

Marvin: So it sounds like nobody misses anything and has to go back out to the property.

Matt: One of the major benefits and one of the key triggers for us in terms of implementing computer vision was to limit re inspections. So this is a pretty heavy demand on the borrower to spend 45 minutes to an hour with somebody in their property

Matt: and we want to limit re inspections as one of the key benefits to the borrower. The big benefit for us is the accuracy in the result of the inspection and being able to drive efficiency and ensuring that we’ve answered everything. In the time that we’ve allotted to the inspection, the human eye doesn’t always see everything in terms of small cracks in the, in the tile or accurately reporting on improvements and all of that comes through the computer vision so that we make the right decision and we produce the right product downstream.

Marvin: Matt, you mentioned valuation modernization, which is a big initiative with the is to make sure that we take appraisal and valuation into the next phase and really help to scale the industry. Can you talk briefly about how computer vision technology supports the valuation modernization initiatives?

Matt: Yes, so our property data inspection is approved by both Fannie Mae and Freddie Mac to be utilized in inspection based waivers and also in hybrid appraisals. So this is delivered downstream to the appraiser in the hybrid appraisals, and the key factor is delivering as accurate an inspection as you possibly can, in order to make that decision.

Matt: But this is a major component in valuation modernization in terms of bifurcating the overall process. And we will automatically add scale with the VALIDITY Pro the property data inspection that supports the valuation modernization and bifurcating it and having the appraisal be able to be done by the desk.

Matt: By desktop appraiser, that’s well adding scale. And in my opinion, adding a lot of accuracy and standardization to the inspection.

Marvin: I guess when I take a step back and look at this valuation modernization initiative that both GSEs, Fannie Mae and Freddie Mac are both involved in pioneering, it really is more than just having a lot of AMCs to do the work, you have to have an appraisal management company that really incorporates the technology, the supporting technology throughout. So that to me sounds like a huge benefit there. How do lenders benefit from SVI’s use of computer vision? We’ve talked about SVI and the appraiser, the homeowner, but how do lenders really benefit?

Matt: Lenders benefit in a couple of ways. One, they benefit in the accuracy of the inspection. So they can trust that with, the human inspector, trained, the application that is really world class in terms of its thoroughness and what it collects, implementing the computer vision into it takes it one step further and ensures that we have captured everything in the property that we can capture and that we’re accurately reporting on the condition, we’re accurately reporting on the improvements and what we see in the property, and we’ve driven a lot of that into our inspection application and a lot of that into our operational processes. In terms of computer vision and how it benefits a lender, one is accuracy. Two, and probably one of the most important, improvements and benefits is protecting against bias.

Matt: One of the major implementations of computer vision inside of the Validity Pro application is being able to blur areas that would otherwise impact bias in the inspection. So, religious artifacts photos, people, you can blur all of that out and you can deliver a factual, unbiased, objective inspection of the property

Matt: Further on, as part of our quality control process, it can help compare what we see in the comparables in the appraisal, or any of our valuations, that can compare those images that you get in MLS to the images that are presented in that appraisal and ensure that you’re comparing condition to condition and that you haven’t left any stone unturned in terms of proper adjustments for that condition, and it gives you another way of trying to validate the comparables, which is critically important.

Matt: So it takes a lot of data that can view a lot more photos than you can as a human. And it can give you a really good tool to focus your energy and attention on where you should to improve the concerns of bias in the valuation industry.

Marvin: Matt, thanks so much for spending time with us today. And we’ll have some more information coming out shortly about the technology and VALIDITY Pro and how that can be used. Thanks for being on the show.

Matt: Thanks, Marvin. Take care.

Ready to get started with our team of experts?


Get in touch with us today to speak to one of our real estate valuation experts and learn how SVI can help your organization succeed.