Artificial Intelligence and Machine learning in practice for car damage detection.

HailMaster Application is an innovative tool that uses augmented reality and computer vision to precisely detect and measure the damaged area with dents left on a car by the hail. While more than half of such car inspection measurements in Europe are still effectuated manually, you can just install the application, take a photo of the damaged area, and wait less than a minute while the HailMaster does all the work for you. The solution was successfully presented at the Automechanika International Trade Fair 2016 in Frankfurt and has heightened the interest of the Hail Dents Repair professionals.

This summer, the first Hail Dents Repair centers in Germany and several other EU countries participated in HailMaster Public Beta Testing what allowed them having a closer look at the technology by using it within the process of hail damage assessment. At the same time, we received valuable practical insights and feedback from the established sphere experts on the possible improvements of our innovation with regards to their trial experience.

The given solution is grounded on the Machine Learning and Artificial Intelligence (AI) technologies, which enjoy the momentum of high demand and rapid development. AI for mobile devices has received a particularly powerful push for advancement recently with Apple having presented its Machine Learning framework called Core ML for the developers. Hence, we see a huge potential of the technologies like HailMaster Application for damage capturing by the use of computer vision not only in automotive industry but also in other businesses where the prompt and efficient defects’ and distinct differences’ recognition may be of demand.

Having spent a while on investigating the most recent approaches to resolving this issue and testing our tool, we have elaborated the most efficient solutions that showed impressive results even in the most difficult hail damage identification cases due to the use of SVM (support vector machine), gabor wavelets, and neuron networks such as CNN (convolutional neural networks).

The updated version is already being tested by our team and proves the successful integration of the latest Machine Learning algorithms and the collected hail damage data. So, the Public Beta version will be available soon for the broader trial, while we will be continuing our research to share the results with you. In case you have any questions or comments, feel free to contact us.

Join us in this great journey!

Example of Hail Dents Detection process using ML in real-time you can watch here:

www.soft-edge.com