Yun-Hao Yuan

A unified multiset canonical correlation analysis framework based on graph embedding for multiple feature extraction

Abstract Multiset canonical correlation analysis (MCCA) can simultaneously reduce the dimensionality of multimodal data. Thus, MCCA is very much suitable and powerful for multiple feature extraction. However, most existing MCCA-related methods are unsupervised algorithms, which are not very effective for pattern classification tasks. In order to improve discriminative power for handling multimodal data, we, in this paper, propose a unified multiset canonical correlation analysis framework based on graph embedding for dimensionality reduction (GbMCC-DR). Under GbMCC-DR framework, three novel supervised multiple feature extraction methods, i.e., GbMCC-LDA, GbMCC-LDE, and GbMCC-MFA are presented by incorporating several well-known graphs. These three methods consider not only geometric structure of multimodal data but also separability of different classes. Moreover, theoretical analysis further shows that, in some specific circumstances, several existing MCCA-related algorithms can be unified into GbMCC-DR framework. Therefore, this proposed framework has good expansibility and generalization. The experimental results on both synthetic data and several popular real-world datasets demonstrate that three proposed algorithms achieve better recognition performance than existing related algorithms, which is also the evidence for effectiveness of GbMCC-DR framework.

Multi-view Latent Hashing for Efficient Multimedia Search

Hashing techniques have attracted broad research interests in recent multimedia studies. However, most of existing hashing methods focus on learning binary codes from data with only one single view, and thus cannot fully utilize the rich information from multiple views of data. In this paper, we propose a novel unsupervised hashing approach, dubbed multi-view latent hashing (MVLH), to effectively incorporate multi-view data into hash code learning. Specifically, the binary codes are learned by the latent factors shared by multiple views from an unified kernel feature space, where the weights of different views are adaptively learned according to the reconstruction error with each view. We then propose to solve the associate optimization problem with an efficient alternating algorithm. To obtain high-quality binary codes, we provide a novel scheme to directly learn the codes without resorting to continuous relaxations, where each bit is efficiently computed in a closed form. We evaluate the proposed method on several large-scale datasets and the results demonstrate the superiority of our method over several other state-of-the-art methods.

Semi-Paired Discrete Hashing: Learning Latent Hash Codes for Semi-Paired Cross-View Retrieval

Due to the significant reduction in computational cost and storage, hashing techniques have gained increasing interests in facilitating large-scale cross-view retrieval tasks. Most cross-view hashing methods are developed by assuming that data from different views are well paired, e.g., text-image pairs. In real-world applications, however, this fully-paired multiview setting may not be practical. The more practical yet challenging semi-paired cross-view retrieval problem, where pairwise correspondences are only partially provided, has less been studied. In this paper, we propose an unsupervised hashing method for semi-paired cross-view retrieval, dubbed semi-paired discrete hashing (SPDH). In specific, SPDH explores the underlying structure of the constructed common latent subspace, where both paired and unpaired samples are well aligned. To effectively preserve the similarities of semi-paired data in the latent subspace, we construct the cross-view similarity graph with the help of anchor data pairs. SPDH jointly learns the latent features and hash codes with a factorization-based coding scheme. For the formulated objective function, we devise an efficient alternating optimization algorithm, where the key binary code learning problem is solved in a bit-by-bit manner with each bit generated with a closed-form solution. The proposed method is extensively evaluated on four benchmark datasets with both fully-paired and semi-paired settings and the results demonstrate the superiority of SPDH over several other state-of-the-art methods in term of both accuracy and scalability.

Laplacian multiset canonical correlations for multiview feature extraction and image recognition

Multiset canonical correlation analysis (MCCA) aims at revealing the linear correlations among multiple sets of high-dimensional data. Therefore, it is only a linear multiview dimensionality reduction technique and such a linear model is insufficient to discover the nonlinear correlation information hidden in multiview data. In this paper, we incorporate the local structure information into MCCA and propose a novel algorithm for multiview dimensionality reduction, called Laplacian multiset canonical correlations (LapMCCs), which simultaneously considers local within-view and local between-view correlations by using nearest neighbor graphs. This makes LapMCC capable of discovering the nonlinear correlation information among multiview data by combining many locally linear problems together. Moreover, we also develop an orthogonal version of LapMCC to preserve the metric structure. The proposed LapMCC method is applied to face and object image recognition. The experimental results on AR, Yale-B, AT&T, and ETH-80 databases demonstrate the superior performance of LapMCC compared to existing multiview dimensionality reduction methods.

A GM-PHD algorithm for multiple target tracking based on false alarm detection with irregular window

Probability hypothesis density (PHD) filter is a suboptimal Bayesian multi-target filter based on random finite set. Gaussian mixture is an approximation scheme to obtain the closed solution of the PHD filter, which is only suitable for linear Gaussian case. However, when targets are moving closely to each other, GM-PHD filter cannot correctly estimate the number of targets and their states. Especially, the estimation accuracy of both target number and their states is rather difficult when targets born and disappear in closely spaced target tracking scenarios. To solve these problems, a novel multiple target tracking algorithm is proposed in this paper. For one hand, when the targets are close, a novel weight redistribution scheme of targets is proposed, which can appropriately modify the weights of the closely spaced targets so that the higher precision of state estimates can be obtained. On the other hand, we propose a false alarm detection method by using an irregular window, in which the multi-scan measurement information is considered to reduce the disturbance of clutter. In numerical experiments, the results demonstrate that the proposed approach can achieve better performance compared to the other existing methods. A GM-PHD algorithm based on false alarm detection with irregular window is proposed.The weight correction of closely spaced targets can be solved by the novel weight redistribution scheme.An irregular window scheme is adopted to solve the false alarm problem in closely spaced target tracking scenario.An improved pruning and merging scheme of Gaussian components can avoid the phenomenon of merging erroneously of different targets.Simulation results show that the proposed algorithm can effectively estimate the number of targets and their states.

Collaborative probabilistic labels for face recognition from single sample per person

Single sample per person (SSPP) recognition is one of the most challenging problems in face recognition (FR) due to the lack of information to predict the variations in the query sample. To address this problem, we propose in this paper a novel face recognition algorithm based on a robust collaborative representation (CR) and probabilisticgraph model, which is called Collaborative Probabilistic Labels (CPL). First, by utilizing label propagation, we construct probabilistic labels for the samples in the generic training set corresponding to those in the gallery set, thus the discriminative information of the unlabeled data can be effectively explored in our method. Then, the adaptive variation type for a given test sample is automatically estimated. Finally, we propose a novel reconstruction-based classifier for the test sample with its corresponding adaptive dictionary and probabilistic labels. The proposed probabilistic graph based model is adaptively robust to various variations in face images, including illumination, expression, occlusion, pose, etc., and is able to reduce required training images to one sample per class. Experimental results on five widely used face databases are presented to demonstrate the efficacy of the proposed approach. Constructed probabilistic graph propagates discrimination from generic to gallery.The adaptive variation type for a given sample can be automatically estimated.CPL incorporates a novel probabilistic label reconstruction based.

Semi-paired hashing for cross-view retrieval

Abstract Hashing techniques have been widely applied in the large-scale cross-view retrieval tasks due to the significant advantage of hash codes in computation and storage efficiency. Most existing cross-view hashing methods can only handle fully-paired scenarios, where all samples from different views are paired. However, such full pairwise correspondences may not be available in practical applications. In this paper, we propose a novel hashing method, named semi-paired hashing (SPH), to deal with a more challenging cross-view retrieval task, where only partial pairwise correspondences are provided in advance. Specifically, SPH aims to preserve within-view similarity and cross-view correlation among multi-view data. Similarity structure within each view is obtained via anchor graph. As limited samples are paired, correlation between unpaired samples is exploited via a simple yet effective approach, which estimates cross-view correlation by partial cross-view pairwise information and within-view similarity structure. Besides, we further incorporate two regression terms between original features and target binary codes to reduce the quantization loss. An efficient iterative algorithm is presented to simultaneously solve hash functions and binary codes. Extensive experiments on two benchmark datasets demonstrate the superiority of SPH over the state-of-the-art methods, especially in the semi-paired scenarios.

Robust Cross-view Hashing for Multimedia Retrieval

Hashing techniques have been widely applied to large-scale cross-view retrieval tasks due to the significant advantage of binary codes in computation and storage efficiency. However, most existing cross-view hashing methods learn binary codes with continuous relaxations, which cause large quantization loss across views. To address this problem, in this letter, we propose a novel cross-view hashing method, where a common Hamming space is learned such that binary codes from different views are consistent and comparable. The quantization loss across views is explicitly reduced by two carefully designed regression terms from original spaces to the Hamming space. In our method, the l2,1-norm regularization is further exploited for discriminative feature selection. To obtain high-quality binary codes, we propose to jointly learn the codes and hash functions, for which an efficient iterative algorithm is presented. We evaluate the proposed method, dubbed Robust Cross-view Hashing (RCH), on two benchmark datasets and the results demonstrate the superiority of RCH over many other state-of-the-art methods in terms of retrieval performance and cross-view consistency.

Multiple extended target tracking algorithm based on GM-PHD filter and spectral clustering

With the increase of the resolution of modern radars and other detection equipments, one target may produce more than one measurement. Such targets are referred to as extended targets. Recently, multiple extended target tracking (METT) has drawn a considerable attention. However, one crucial problem is how to partition the measurement sets accurately and rapidly. In this paper, an improved METT algorithm is proposed based on the Gaussian mixture probability hypothesis density (GM-PHD) filter and an effective partition method using spectral clustering technique. First, the density analysis technique is introduced to eliminate the disturbance of clutter, and then the spectral clustering technique based on neighbor propagation is used to partition the measurements. Finally, the GM-PHD filter is implemented to achieve the METT. Simulation results show that the proposed algorithm has a better performance, especially a better real-time performance, than the conventional distance partition and K-means++ methods.

Short text clustering based on Pitman-Yor process mixture model

For finding the appropriate number of clusters in short text clustering, models based on Dirichlet Multinomial Mixture (DMM) require the maximum possible cluster number before inferring the real number of clusters. However, it is difficult to choose a proper number as we do not know the true number of clusters in short texts beforehand. The cluster distribution in DMM based on Dirichlet process as prior goes down exponentially as the number of clusters increases. Therefore, we propose a novel model based on Pitman-Yor Process to capture the power-law phenomenon of the cluster distribution in the paper. Specifically, each text chooses one of the active clusters or a new cluster with probabilities derived from the Pitman-Yor Process Mixture model (PYPM). Discriminative words and nondiscriminative words are identified automatically to help enhance text clustering. Parameters are estimated efficiently by collapsed Gibbs sampling and experimental results show PYPM is robust and effective comparing with the state-of-the-art models.