Jun Gao

Cooperative Sparse Representation in Two Opposite Directions for Semi-Supervised Image Annotation

Recent studies have shown that sparse representation (SR) can deal well with many computer vision problems, and its kernel version has powerful classification capability. In this paper, we address the application of a cooperative SR in semi-supervised image annotation which can increase the amount of labeled images for further use in training image classifiers. Given a set of labeled (training) images and a set of unlabeled (test) images, the usual SR method, which we call forward SR, is used to represent each unlabeled image with several labeled ones, and then to annotate the unlabeled image according to the annotations of these labeled ones. However, to the best of our knowledge, the SR method in an opposite direction, that we call backward SR to represent each labeled image with several unlabeled images and then to annotate any unlabeled image according to the annotations of the labeled images which the unlabeled image is selected by the backward SR to represent, has not been addressed so far. In this paper, we explore how much the backward SR can contribute to image annotation, and be complementary to the forward SR. The co-training, which has been proved to be a semi-supervised method improving each other only if two classifiers are relatively independent, is then adopted to testify this complementary nature between two SRs in opposite directions. Finally, the co-training of two SRs in kernel space builds a cooperative kernel sparse representation (Co-KSR) method for image annotation. Experimental results and analyses show that two KSRs in opposite directions are complementary, and Co-KSR improves considerably over either of them with an image annotation performance better than other state-of-the-art semi-supervised classifiers such as transductive support vector machine, local and global consistency, and Gaussian fields and harmonic functions. Comparative experiments with a nonsparse solution are also performed to show that the sparsity plays an important role in the cooperation of image representations in two opposite directions. This paper extends the application of SR in image annotation and retrieval.

Optimizing widths with PSO for center selection of Gaussian radial basis function networks

The radial basis function (RBF) centers play different roles in determining the classification capability of a Gaussian radial basis function neural network (GRBFNN) and should hold different width values. However, it is very hard and time-consuming to optimize the centers and widths at the same time. In this paper, we introduce a new insight into this problem. We explore the impact of the definition of widths on the selection of the centers, propose an optimization algorithm of the RBF widths in order to select proper centers from the center candidate pool, and improve the classification performance of the GRBFNN. The design of the objective function of the optimization algorithm is based on the local mapping capability of each Gaussian RBF. Further, in the design of the objective function, we also handle the imbalanced problem which may occur even when different local regions have the same number of examples. Finally, the recursive orthogonal least square (ROLS) and genetic algorithm (GA), which are usually adopted to optimize the RBF centers, are separately used to select the centers from the center candidates with the initialized widths, in order to testify the validity of our proposed width initialization strategy on the selection of centers. Our experimental results show that, compared with the heuristic width setting method, the width optimization strategy makes the selected centers more appropriate, and improves the classification performance of the GRBFNN. Moreover, the GRBFNN constructed by our method can attain better classification performance than the RBF LS-SVM, which is a state-of-the-art classifier.

Generic object recognition with regional statistical models and layer joint boosting

This paper presents novel regional statistical models for extracting object features, and an improved discriminative learning method, called as layer joint boosting, for generic multi-class object detection and categorization in cluttered scenes. Regional statistical properties on intensities are used to find sharing degrees among features in order to recognize generic objects efficiently. Based on boosting for multi-classification, the layer characteristic and two typical weights in sharing-code maps are taken into account to keep the maximum Hamming distance in categories, and heuristic search strategies are provided in the recognition process. Experimental results reveal that, compared with interest point detectors in representation and multi-boost in learning, joint layer boosting with statistical feature extraction can enhance the recognition rate consistently, with a similar detection rate.

Sequential Pattern Mining with Wildcards

Sequential pattern mining is an important research task in many domains, such as biological science. In this paper, we study the problem of mining frequent patterns from sequences with wildcards. The user can specify the gap constraints with flexibility. Given a subject sequence, a minimal support threshold and a gap constraint, we aim to find frequent patterns whose supports in the sequence are no less than the given support threshold. We design an efficient mining algorithm MAIL that utilizes the candidate occurrences of the prefix to compute the support of a pattern that avoids the rescanning of the sequence. We present two pruning strategies to improve the completeness and the time efficiency of MAIL. Experiments show that MAIL mines 2 times more patterns than one of its peers and the time performance is 12 times faster on average than its another peer.

A BIT-PARALLEL ALGORITHM FOR SEQUENTIAL PATTERN MATCHING WITH WILDCARDS

Pattern matching with both gap constraints and the one-off condition is a challenging topic, especially in bioinformatics, information retrieval, and dictionary query. Among the algorithms to solve the problem, the most efficient one is SAIL, which is time consuming, especially when the pattern is long. In addition, existing algorithms based on bit-parallelism cannot handle a pattern that has only one pattern character between successive wildcards and the minimum local length constraints are zero. We propose an algorithm BPBM to handle online sequential pattern matching. In BPBM, an extended bit-parallelism operation is used to accelerate the matching process. An effective transition window mechanism with two nondeterministic finite state automatons (NFAs) is adopted to drop the useless scan window. It identifies gap constraints automatically and just scans once to export occurrences with exact match positions. Theoretical analysis and experimental results show that the BPBM algorithm is more competitive than other peers. It has an absolute advantage on search time complexity. It also has better stability that decreases operation costs with the increasing of the size of sequence alphabet or the length of the pattern. We also study off-line pattern matching. With twice pruning, left-most and right-most, we can increase the matching ratio about 2.08% on average.

Pattern matching with wildcards based on key character location

Pattern matching with wildcards is a complex problem and this problem has wide potential application in text search, biological sequences and information security etc. We propose a new algorithm called Quicksearch, for pattern matching with wildcards and length constraints based on key character location and subspace partition. This new algorithm increases by 40%–60% searching efficiency in comparison with SAIL when characters of pattern P in text T are unevenly distributed..

MAIL: mining sequential patterns with wildcards

Sequential pattern mining is an important research task in many domains, such as biological science. In this paper, we study the problem of mining frequent patterns from sequences with wildcards. The user can specify the gap constraints with flexibility. Given a subject sequence, a minimal support threshold and a gap constraint, we aim to find frequent patterns whose supports in the sequence are no less than the given support threshold. We design an efficient mining algorithm MAIL. Two pattern growth strategies are proposed to improve the completeness and the time efficiency. One is based on the candidate occurrence pruning, and the other uses an occurrence graph. A random data generator is designed to test the completeness on artificial data. Experiments on DNA sequences show that MAIL mines four times more patterns than one of its peers and the time performance is six times faster on average than its another peer. We also give a concrete example in which our algorithm is applied on DNA sequences to find interesting patterns.

Image set classification based on cooperative sparse representation

Abstract Image set classification has been widely applied to many real-life scenarios including surveillance videos, multi-view camera networks and personal albums. Compared with single image based classification, it is more promising and therefore has attracted significant research attention in recent years. Traditional (forward) sparse representation (fSR) just makes use of training images to represent query ones. If we can find complementary information from backward sparse representation (bSR) which represents query images with training ones, the performance will be likely to be improved. However, for image set classification, the way to produce additional bases for bSR is a problem concerned as there is no other bases than the query set itself. In this paper, we extend cooperative sparse representation (CoSR) method, which integrates fSR and bSR together, to image set classification. In this process, we propose two schemes, namely ‘Learning Bases’ and ‘Training Sets Division’, to produce the additional dictionary for bSR. And different from previous work, our work considers scene classification as a problem of image set classification, which will provide a new insight for scene classification. Experimental results show that the proposed model can obtain competitive recognition rates for image set classification. By combining information from these two opposite SRs, better results can be achieved. Also the feasibility for the formulation of image set classification on scene classification is validated.

Mining Frequent Patterns with Gaps and One-Off Condition

Mining frequent patterns with a gap requirement from sequences is an important step in many domains, such as biological sciences. Given a character sequence S of length L, a certain threshold and a gap constraint, we aim to discover frequent patterns whose supports in S are no less than the given threshold value. A frequent pattern P can have wildcards, and the numbers of the wildcards between elements of P must fulfill user-specified gap constraints. Also, this mining process satisfies the one-off condition and an Apriori-like property to be efficient. Experiments show that our method can mine as many frequent patterns with wildcards as the existing MPP algorithm, but has a much better performance in time.

Beyond shape: incorporating color invariance into a biologically inspired feedforward model of category recognition

Being lack of theoretical support from biological cues in computer vision, current computational and learning approaches of object categorization mostly aim at better performances neglecting analysis on framework in human brain for visual information processing materially which cause little-marginal improvement and more complexity. Focusing on the uncertainty of color mechanism in visual cortex and motivating from biological issues on shape information, we present the model incorporating color invariant descriptors and plausible shape feature biologically to formulate the robust representation of each category with only simple SVM classifier to achieve the amazing performance. Our model has the characteristics of illumination, scale, position, orientation, viewpoint invariance, and competitive with current algorithms on only a few training examples from several data sets, including Caltech 101 and GRAZ for category recognition. Also, experimental results show the robustness when challenged by noisy or blurred images.