Published Paper Details:

This study presents a novel approach utilizing deep learning techniques to develop an automated system for the detection of automotive damage and subsequent prediction of repair costs. The system comprises two primary components, namely a model for detecting damage and a model for predicting prices. The damage identification model utilizes a convolutional neural network architecture, specifically VGG16 and Resnet50, which have been trained to accurately classify and recognize various forms of damage present in car photos. The price prediction model utilized in this study is an XGBoost model, which has been trained to forecast the market value of a car by considering its condition as a key factor. The evaluation of the system was conducted using a dataset comprising more than 10,000 photographs of vehicles. The findings illustrate the capacity of deep learning techniques in automating the detection of automotive damage and predicting its associated pricing. The system under consideration has the potential to optimize the efficiency of insurance claim processing, used automobile appraisal, and maintenance planning procedures.