A Car Face Parts Detection Algorithm Based on Faster R-CNN
Zhihao Zhou, Xiying Li, Mingkai Qiu · CICTP 2018 · 2018
The main appearance difference between different types of vehicles is located in the front face area, so the car face parts detection is a key role in fine-grained vehicle recognition. This paper presents a faster R-CNN-based method to detect the position and identify each part of vehicle front face in a complex environment. First, the object information is carried out by K-means clustering and the feature is extracted by VGG-16 network. Second, the candidate regions are obtained by region proposal network (RPN), then uses the Fast R-CNN to obtain the categories and location information of vehicle front face parts. In this paper, 4199 vehicle images of CompCars dataset were used for network training and testing. The experimental results show that the average IoU of the vehicle front face parts is 74.97% and the average recognition precision is 89.59%. Compared with other object detection algorithms, the proposed algorithm shows excellent performance in detecting ability and recognition effect.