Turning pictures of broken autos into correct claims data is important for insurance coverage companies. For that cause, KASIKORN Enterprise-Know-how Group (KBTG) Labs, a number one tech firm in Thailand, is utilizing Machine studying (ML) expertise to assist velocity up this course of, and make sure the accuracy of each declare.
KBTG labored with Muang Thai Insurance coverage PCL (MTI), an insurance coverage firm in Thailand that gives fireplace, marine, and vehicle insurance coverage. When a car is broken in an accident the restore store usually assesses the injury and submits a declare to the insurance coverage firm for processing. These claims are validated by the insurance coverage firm after which processed with intermediaries that validate the claims and submit images as proof.
At present, this whole means of assessing the injury and evaluating the declare is totally handbook, which slows down processing time and is dear, as a result of a number of companions and intermediaries are concerned. To deal with this subject, KBTG Labs labored with Google Cloud to plan an answer to this drawback.
Constructing an answer structure
The present dataset used for claims processing had many information high quality and amount challenges, For instance, some pictures may be taken from too far a distance or too near establish the particular space of injury, or the descriptions could also be incorrect.