Vision-based vehicle detection is the most basic and important technology in advanced driver\nassistance systems. In this paper, we propose a vehicle detection framework using selective multi-stage\nfeatures in convolutional neural networks (CNNs) to improve vehicle detection performance. A 10-layer\nCNN model was designed and visualization techniques were used to selectively extract features from the\nactivation feature map, called selective multi-stage features. The proposed features contain characteristic\nvehicle image information and are more robust than traditional features against noise. We trained the\nAdaBoost algorithm using these features to implement a vehicle detector. The experimental results verified\nthat the proposed vehicle detection framework exhibited better performance than previous frameworks.
Loading....