Yongbin Gao

Vehicle Make Recognition Based on Convolutional Neural Network

Vehicle analysis has been investigated for decades, which involves license plate recognition, intelligent traffic. Among these applications, vehicle make recognition is a challenging task due to the close appearance between car models. In this paper, we propose an architecture to recognize vehicle make based on convolutional neural network (CNN). The moving car is first localized by frame difference, the resultant binary image is used to detect the frontal view of a car by a symmetry filter. The detected frontal view is used to train and test the CNN. Experimental results show that our proposed framework achieves favorable recognition accuracy.

Local Tiled Deep Networks for Recognition of Vehicle Make and Model

Vehicle analysis involves license-plate recognition (LPR), vehicle-type classification (VTC), and vehicle make and model recognition (MMR). Among these tasks, MMR plays an important complementary role in respect to LPR. In this paper, we propose a novel framework for MMR using local tiled deep networks. The frontal views of vehicle images are first extracted and fed into the local tiled deep networks for training and testing. A local tiled convolutional neural network (LTCNN) is proposed to alter the weight sharing scheme of CNN with local tiled structure. The LTCNN unties the weights of adjacent units and then ties the units k steps from each other within a local map. This architecture provides the translational, rotational, and scale invariance as well as locality. In addition, to further deal with the colour and illumination variation, we applied the histogram oriented gradient (HOG) to the frontal view of images prior to the LTCNN. The experimental results show that our LTCNN framework achieved a 98% accuracy rate in terms of vehicle MMR.

Viewpoint Unconstrained Face Recognition Based on Affine Local Descriptors and Probabilistic Similarity

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we propose using the combination of Affine Scale Invariant Feature Transform (SIFT) and Probabilistic Similarity for face recognition under a large viewpoint change. Affine SIFT is an extension of SIFT algorithm to detect affine invariant local descriptors. Affine SIFT generates a series of different viewpoints using affine transformation. In this way, it allows for a viewpoint difference between the gallery face and probe face. However, the human face is not planar as it contains significant 3D depth. Affine SIFT does not work well for significant change in pose. To complement this, we combined it with probabilistic similarity, which gets the log likelihood between the probe and gallery face based on sum of squared difference (SSD) distribution in an offline learning process. Our experiment results show that our framework achieves impressive better recognition accuracy than other algorithms compared on the FERET database.

Car manufacturer and model recognition based on scale invariant feature transform

Vehicle analysis involves licence plate recognition, vehicle type recognition, and car manufacturer and model recognition. Car manufacturer and model recognition plays an important role in providing supplementary information to licence plate recognition for the unique identification of a car. In this paper, we propose a framework to recognition car manufacturer and its model based on scale invariant feature transform (SIFT). We first detect a moving car using frame differences; the resultant binary image is used to detect the frontal view of a car by a symmetry filter. The detected frontal view is then used to identify a car based on SIFT algorithm. Experimental results show that our proposed framework achieves favourable recognition accuracy.