Cuihua Li

Discriminative Object Tracking via Sparse Representation and Online Dictionary Learning

We propose a robust tracking algorithm based on local sparse coding with discriminative dictionary learning and new keypoint matching schema. This algorithm consists of two parts: the local sparse coding with online updated discriminative dictionary for tracking (SOD part), and the keypoint matching refinement for enhancing the tracking performance (KP part). In the SOD part, the local image patches of the target object and background are represented by their sparse codes using an over-complete discriminative dictionary. Such discriminative dictionary, which encodes the information of both the foreground and the background, may provide more discriminative power. Furthermore, in order to adapt the dictionary to the variation of the foreground and background during the tracking, an online learning method is employed to update the dictionary. The KP part utilizes refined keypoint matching schema to improve the performance of the SOD. With the help of sparse representation and online updated discriminative dictionary, the KP part are more robust than the traditional method to reject the incorrect matches and eliminate the outliers. The proposed method is embedded into a Bayesian inference framework for visual tracking. Experimental results on several challenging video sequences demonstrate the effectiveness and robustness of our approach.

Unifying visual saliency with HOG feature learning for traffic sign detection

Traffic sign detection is important to a robotic vehicle that automatically drives on roads. In this paper, an efficient novel approach which is enlighten by the process of the human vision is proposed to achieve automatic traffic sign detection. The detection method combines bottom-up traffic sign saliency region with learning based top-down features of traffic sign guided search. The bottom-up stage could obtain saliency region of traffic sign and achieve computational parsimony using improved Model of Saliency-Based Visual Attention. The top-down stage searches traffic sign in these traffic sign saliency regions based on the feature Histogram of Oriented Gradient (HOG) and the classifier Support Vector Mechine (SVM). Experimental results show that, the proposed approach can achieve robustness to illumination, scale, pose, viewpoint change and even partial occlusion. The samllest detection size of traffic sign is 14×14, the average detection rate is 98.3% and the false positive rate is 5.09% in test image data set.

Coupled Deep Autoencoder for Single Image Super-Resolution

Sparse coding has been widely applied to learning-based single image super-resolution (SR) and has obtained promising performance by jointly learning effective representations for low-resolution (LR) and high-resolution (HR) image patch pairs. However, the resulting HR images often suffer from ringing, jaggy, and blurring artifacts due to the strong yet ad hoc assumptions that the LR image patch representation is equal to, is linear with, lies on a manifold similar to, or has the same support set as the corresponding HR image patch representation. Motivated by the success of deep learning, we develop a data-driven model coupled deep autoencoder (CDA) for single image SR. CDA is based on a new deep architecture and has high representational capability. CDA simultaneously learns the intrinsic representations of LR and HR image patches and a big-data-driven function that precisely maps these LR representations to their corresponding HR representations. Extensive experimentation demonstrates the superior effectiveness and efficiency of CDA for single image SR compared to other state-of-the-art methods on Set5 and Set14 datasets.

Image clustering by hyper-graph regularized non-negative matrix factorization

Abstract Image clustering is a critical step for the applications of content-based image retrieval, image annotation and other high-level image processing. To achieve these tasks, it is essential to obtain proper representation of the images. Non-negative Matrix Factorization (NMF) learns a part-based representation of the data, which is in accordance with how the brain recognizes objects. Due to its psychological and physiological interpretation, NMF has been successfully applied in a wide range of application such as pattern recognition, image processing and computer vision. On the other hand, manifold learning methods discover intrinsic geometrical structure of the high dimension data space. Incorporating manifold regularizer to standard NMF framework leads to novel performance. In this paper, we proposed a novel algorithm, call Hyper-graph regularized Non-negative Matrix Factorization (HNMF) for this purpose. HNMF captures intrinsic geometrical structure by constructing a hyper-graph instead of a simple graph. Hyper-graph model considers high-order relationship of samples and outperforms simple graph model. Empirical experiments demonstrate the effectiveness of the proposed algorithm in comparison to the state-of-the-art algorithms, especially some related works based on NMF.

Low-rank matrix factorization with multiple Hypergraph regularizer

This paper presents a novel low-rank matrix factorization method, named MultiHMMF, which incorporates multiple Hypergraph manifold regularization to the low-rank matrix factorization. In order to effectively exploit high order information among the data samples, the Hypergraph is introduced to model the local structure of the intrinsic manifold. Specifically, multiple Hypergraph regularization terms are separately constructed to consider the local invariance; the optimal intrinsic manifold is constructed by linearly combining multiple Hypergraph manifolds. Then, the regularization term is incorporated into a truncated singular value decomposition framework resulting in a unified objective function so that matrix factorization is changed into an optimization problem. Alternating optimization is used to solve the optimization problem, with the result that the low dimensional representation of data space is obtained. The experimental results of image clustering demonstrate that the proposed method outperforms state-of-the-art data representation methods. We propose a novel low rank matrix factorization method.We incorporate multiple Hypergraph manifold regularization to the matrix factorization.We adopt alternating optimization to solve the optimization problem.

Online multiple instance gradient feature selection for robust visual tracking

In this paper, we focus on learning an adaptive appearance model robustly and effectively for object tracking. There are two important factors to affect object tracking, the one is how to represent the object using a discriminative appearance model, the other is how to update appearance model in an appropriate manner. In this paper, following the state-of-the-art tracking techniques which treat object tracking as a binary classification problem, we firstly employ a new gradient-based Histogram of Oriented Gradient (HOG) feature selection mechanism under Multiple Instance Learning (MIL) framework for constructing target appearance model, and then propose a novel optimization scheme to update such appearance model robustly. This is an unified framework that not only provides an efficient way of selecting the discriminative feature set which forms a powerful appearance model, but also updates appearance model in online MIL Boost manner which could achieve robust tracking overcoming the drifting problem. Experiments on several challenging video sequences demonstrate the effectiveness and robustness of our proposal.

An effective unconstrained correlation filter and its kernelization for face recognition

In this paper, an effective unconstrained correlation filter called Unconstrained Optimal Origin Tradeoff Filter (UOOTF) is presented and applied to robust face recognition. Compared with the conventional correlation filters in Class-dependence Feature Analysis (CFA), UOOTF improves the overall performance for unseen patterns by removing the hard constraints on the origin correlation outputs during the filter design. To handle non-linearly separable distributions between different classes, we further develop a non-linear extension of UOOTF based on the kernel technique. The kernel extension of UOOTF allows for higher flexibility of the decision boundary due to a wider range of non-linearity properties. Experimental results demonstrate the effectiveness of the proposed unconstrained correlation filter and its kernelization in the task of face recognition.

A novel unconstrained correlation filter and its application in face recognition

In this paper, a novel unconstrained correlation filter called Unconstrained Optimal Origin Tradeoff Filter (UOOTF) is presented and applied to face recognition. Compared with the conventional correlation filters in Class-dependence Feature Analysis (CFA), UOOTF increases the overall performance for unseen patterns by removing the hard constraints on the outputs during the filter design. Experimental results on the popular FERET, FRGC and CAS-PEAL R1 face databases show the effectiveness of the proposed unconstrained correlation filter.

Image super-resolution base on multi-kernel regression

In this paper, a novel approach to single image super-resolution based on the multi-kernel regression is presented. This approach focuses on learning the map between the space of high-resolution image patches and the space of blurred high-resolution image patches, which are the interpolation results generated from the corresponding low-resolution images. Kernel regression based super-resolution approaches are promising, but kernel selection is a critical problem. In order to avoid demanding and time-consuming cross validation for kernel selection, we propose multi-kernel regression (MKR) model for image Super-Resolution (SR). Considering the multi-kernel regression model is prohibited when the training data is large-scale, we further propose a prototype MKR algorithm which can reduce the computational complexity. Extensive experimental results demonstrate that our approach is effective and achieves a high quality performance in comparison with other super-resolution methods.

Multi-Frame Image Super-Resolution Based on Regularization Scheme

Super-resolution (SR) reconstruction produces one or a series of high-resolution images from a series of low-resolution images. In this paper, we apply the regularization-based SR image reconstruction method on the basis of multi-frame image SR. Fisrstly, a linear observation model is utilized to associate the recorded LR images with the unknown reconstructed HR image estimates, and we apply the bilateral total variation operator as a regularization term. Moreover, the basic principal of this algorithm is presented, and we thoroughly analyze the selection of the cost-function and the regularization term by comparing of experiments. According to some connective experiments, the algorithm is proved to be effective and robust, and it can better preserve the details of the image.