Car Damage Identification System Using Convolutional Neural Network

The number of individuals driving has increased daily, leading to a rise in the number of auto insurance claims. The lengthy process of registering, processing, and determining the outcome of each claim includes the service engineer's manual assessment, which serves as the basis for the damage report, and the physical examination carried out by an insurance company surveyor. It would be advantageous for both the organization and the client if this entire procedure were automated. This system utilizes images of the car in a damaged state as input. It then generates pertinent data, including the locations of the damaged sections (front, rear, and sides) and an assessment of the extent of the damage (none, minor, or substantial). This system uses the Convolutional Neural Network (CNN) Model to analyze images automatically for auto insurance, providing the user with information on the location and extent of the vehicle damage. ImageNet weights and the VGG16 architecture are also used here. Transfer learning is used for a variety of tasks, such as finding and classifying different kinds of auto parts and issues. The logistic regression classifier is trained using feature and label extraction methods.