Car detection and damage segmentation in the real scene using a deep learning approach

Automatically detecting the outer car surface damage can considerably reduce the cost of processing premium assertion, and provide satisfaction for vehicle users. Since computer vision has a huge development among different research areas during recent years, the utilization of computer vision as a serious branch of science has also affected the object detection field. In this study, we develop an automated car and damage detection method based on a cascade Convolutional Neural Network (CNN). The presented method utilizes a pixel-based approach using two distinct CNNs, to determine the damage in outer region of a car among the achieved images. The experimental results indicate our proposed method obtains high performance in comparison to other state-of-the-art methods.