Can This AI Platform Detect Damage Severity?
Nov. 9, 2020—A tech company called Fyusion developed a photography platform that creates 3D images using a mobile device. The company just announced that it has matched it with artificial intelligence capabilities to assess vehicle damage in the field.
The company's AI is called ALIS, which stands for AI-Based Lightfield Information Suite. It's geared toward 3D imaging and visual understanding, and it can be used in a variety of applications.
The company says that the program will simplify inspection processes for anyone who needs to share damage information. That could include wholesaling, trade-in appraisal, lease-end inspections, and fleet management, according to a press release.
It's apparent that this technology could also benefit collision repair and insurance procedures as well. ADAPT reached Fyusion CEO and co-founder Radu Rusu to learn more about the development and uses for this platform.
What kinds of specialized knowledge was needed to train the ALIS system for this kind of purpose? What was that process like?
We are a technology company with a strong background in robotics, computer vision and machine learning. Fyusion was founded with the idea to develop tools and technology to help better understand the world through 3D technologies. In the last few years, we have been developing 3D scan technology for smartphones used by hundreds of millions of users around the world. 3D scanning technology is the base of our vehicle damage detection, which made developing ALIS a natural step for us. Today we have a team of over twenty PHD researchers focused on making ALIS smarter.
After a vehicle images are into the system, what kinds of damage can the ALIS system detect and analyze?
ALIS is constantly evolving and getting smarter, and currently is able to detect most damages present on the vehicle at a 90% accuracy rate. As part of the damage analysis, ALIS is also able to identify and quantify the area affected by the damage and the exact location of the damage in the vehicle. These are key components to assess the impact of the damage and, in general, the condition of the vehicle.
In what way does a human help to finish out the process?
Once the capture has been uploaded by the user, ALIS automatically analyzes the 360 capture and the damage images, generating the condition report as a result. Right after the condition report is generated, a team of highly skilled humans review the condition report for potential inaccuracies and validate the report before the final release to the customer. This process is a quality assurance process that allows us to ensure the quality of the final report is equivalent to a report generated by a professional appraiser on-site.
Do you envision the ALIS platform being used for other purposes related to automotive repair? What might they be?
Our current focus is to develop the best technology and services to assess the condition of the exterior of the vehicle in a way that is highly accurate, reliable, scalable and cost-efficient for our customers. Assessing the condition of the exterior of a vehicle is a big technology challenge and we believe that our 3D technology will provide results as good or better than professional appraisers. Thanks to our AI-first approach, this same vision can be applied to other use cases related to vehicle repair that will offer a lot of value to our clients. Our next area of focus will be mapping and analyzing the condition of the vehicle interior applying the same principles as for exterior damage analysis.
Image courtesy of Fyusion