Skip to main content

The Auto.ru service has launched a neural network that can determine the condition of a car from pictures. This will allow you to more accurately understand the real value of the car for sale. The development of the algorithm involved the results of an analysis of more than 100,000 cars, which were evaluated by experts. Based on photos of the exterior and interior, the neural network identifies defects in the body and interior of the car, including corrosion and traces of poorly performed repairs.

This will, among other things, allow professional buyers to navigate based on an assessment of the vehicle’s condition. Experts believe that the introduction of a neural network when evaluating a car will be a real revolution in the world of car sales, making it possible to more accurately determine the price of a car. Already now, users of the service can see the car’s rating when viewing ads with photographs that are suitable for analysis. The exception is damaged cars that are not subject to inspection, but such cars are a small number in the total supply.

Earlier, expert Ostwald told which cars are most suitable for a novice driver. Examples of such cars are named in which it will be safer and more comfortable for beginners to travel.

Leave a Reply