How to import image-level labels as weak supervision to direct the region-level labeling task is the core task of weakly-supervised semantic segmentation. In this paper, we focus on designing an effective but simple weakly-supervised constraint, and propose an exclusive constrained discriminative learning model for image semantic segmentation. To be specific, we employ a discriminative linear regression model to assign subsets of superpixels with different labels. During the assignment, we construct an exclusive weakly-supervised constraint term to suppress the labeling responses of each superpixel on the labels outside its parent image-level label set. Besides, a spectral smoothing term is integrated to encourage that both visually and semantically similar superpixels have similar labels. Combining these terms, we formulate the problem as a convex objective function, which can be easily optimized via alternative iterations. Extensive experiments on MSRC-21 and LabelMe datasets demonstrate the effectiveness of the proposed model.

Face recognition has been widely studied due to its importance in various applications. However, the case that both training images and testing images are corrupted is not well addressed. Motivated by the success of low-rank matrix recovery, we propose a novel semi-supervised low-rank matrix recovery algorithm for robust face recognition. The proposed method can learn robust discriminative representations for both training images and testing images simultaneously by exploiting the classwise block-diagonal structure. Specifically, low-rank matrix approximation can handle the possible contamination of data. Moreover, the classwise block-diagonal structure is exploited to promote discrimination of representations for robust recognition. The above issues are formulated into a unified objective function and we design an efficient optimization procedure based on augmented Lagrange multiplier method to solve it. Extensive experiments on three public databases are performed to validate the effectiveness of our approach. The strong identification capability of representations with block-diagonal structure is verified.

Face recognition has been widely studied due to its importance in various applications. However, the case that both training images and testing images are corrupted is not well solved. To address such a problem, this paper proposes a semisupervised learning algorithm for robust face recognition. In particular, we consider three items in the proposed formulation. First, a low-rank and sparse representation for face recognition is required to handle the possible contamination of the whole data. Second, a classwise block-diagonal structure of the learned representation is expected to promote discrimination among different classes. With the structure regularization, we make the samples from different classes be reconstructed with different bases as much as possible. Third, a compact and discriminative dictionary should be learnt to handle the problem of corrupted data. Extensive experiments on three public databases are performed to validate the effectiveness of our approach. The strong identification capability of representation with block-diagonal structure is verified.

Dynamic scene analysis has become a popular research area especially in video surveillance. The goal of this paper is to mine semantic motion patterns and detect abnormalities deviating from normal ones occurring in complex dynamic scenarios. To address this problem, we propose a data-driven and scene-independent approach, namely, Bilayer sparse topic model (BiSTM), where a given surveillance video is represented by a word-document hierarchical generative process. In this BiSTM, motion patterns are treated as latent topics sparsely distributed over low-level motion vectors, whereas a video clip can be sparsely reconstructed by a mixture of topics (motion pattern). In addition to capture the characteristic of extreme imbalance between numerous typical normal activities and few rare abnormalities in surveillance video data, a one-class constraint is directly imposed on the distribution of documents as a discriminant priori. By jointly learning topics and one-class document representation within a discriminative framework, the topic (pattern) space is more specific and explicit. An effective alternative iteration algorithm is presented for the model learning. Experimental results and comparisons on various public data sets demonstrate the promise of the proposed approach.

Effective video presentation and summarization techniques are critical for fast browsing of video content. In this paper, we propose a novel presentation approach to vividly depict the moving process of a specific object in a surveillance video, which aims at effectively summarizing video content by a static image named narrative. Firstly, the object of interest is extracted and segmented from the video to form a spatio-temporal object tube. Then three criteria are proposed to select the most representative objects from this tube. We formulate the object selecting process as an energy minimization problem, in which each energy term measures a corresponding criterion cost. We maximally preserve the changes of appearance and behavior while remove other redundant content as much as possible. Finally, the selected representative objects are stitched to the background image by Poisson editing. Experimental results show the promise of the proposed approach.

With the proliferation of cameras in public areas, it becomes increasingly desirable to develop fully automated surveillance and monitoring systems. In this paper, we propose a novel unsupervised approach to automatically explore motion patterns occurring in dynamic scenes under an improved sparse topical coding (STC) framework. Given an input video with a fixed camera, we first segment the whole video into a sequence of clips (documents) without overlapping. Optical flow features are extracted from each pair of consecutive frames, and quantized into discrete visual words. Then the video is represented by a word-document hierarchical topic model through a generative process. Finally, an improved sparse topical coding approach is proposed for model learning. The semantic motion patterns (latent topics) are learned automatically and each video clip is represented as a weighted summation of these patterns with only a few nonzero coefficients. The proposed approach is purely data-driven and scene independent (not an object-class specific), which make it suitable for very large range of scenarios. Experiments demonstrate that our approach outperforms the state-of-the art technologies in dynamic scene analysis.

We propose an image classification framework by leveraging the non-negative sparse coding, low-rank and sparse matrix decomposition techniques (LR-Sc+SPM). First, we propose a new non-negative sparse coding along with max pooling and spatial pyramid matching method (Sc+SPM) to extract local features' information in order to represent images, where non-negative sparse coding is used to encode local features. Max pooling along with spatial pyramid matching (SPM) is then utilized to get the feature vectors to represent images. Second, motivated by the observation that images of the same class often contain correlated (or common) items and specific (or noisy) items, we propose to leverage the low-rank and sparse matrix recovery technique to decompose the feature vectors of images per class into a low-rank matrix and a sparse error matrix. To incorporate the common and specific attributes into the image representation, we still adopt the idea of sparse coding to recode the Sc+SPM representation of each image. In particular, we collect the columns of the both matrixes as the bases and use the coding parameters as the updated image representation by learning them through the locality-constrained linear coding (LLC). Finally, linear SVM classifier is leveraged for the final classification. Experimental results show that the proposed method achieves or outperforms the state-of-the-art results on several benchmarks.

With the rapid increasing of video cameras, large amount of video data everyday brings the problem of video storage and browsing. In this paper, we propose a novel approach to video reshuffling with a group of static images to effectively summarize the video content. Each static image called narrative is generated to depict the behavior of a specific object or a special event. Firstly background subtraction and object tracking are employed to extract the segmentations of moving objects and corresponding trajectories. After that, we apply three sampling rules to optimized select representative object samples from the spatial-temporal object tube and stitch them to the background image by Poisson blending. Experimental results show the promise of the proposed approach.

The bag-of-visual-words (BoW) representation has received wide application and public acceptance for visual categorization. However, the histogram based image representation ignores the spatial information and correlations among visual words. To tackle these problems, in this paper, we propose to use some image regions called `components', as the higher-level visual elements to represent an image associating with the lower-level elements of `visual words'. Then we formulate the task of visual categorization into two progressive relationships among a given concept and the two-level visual elements of images, i.e., visual-words-to-components and components-to-concept. Firstly, component level linear SVM classifiers are learned to model the relationship between visual words and components, then the output of these SVM classifiers are linearly combined to model the relationships between components and concept. Experiments on the Scene-15 dataset and the Oxford Flowers dataset demonstrate the effectiveness of the proposed method.

In this paper, we propose a novel partwise framework for cross-parameterization between 3D mesh models. Unlike most existing methods that use regular parameterization domains, our framework uses nonregular approximation domains to build the cross-parameterization. Once the nonregular approximation domains are constructed for 3D models, different (and complex) input shapes are transformed into similar (and simple) shapes, thus facilitating the cross-parameterization process. Specifically, a novel nonregular domain, the convex hull, is adopted to build shape correspondence. We first construct convex hulls for each part of the segmented model, and then adopt our convex-hull cross-parameterization method to generate compatible meshes. Our method exploits properties of the convex hull, e.g., good approximation ability and linear convex representation for interior vertices. After building an initial cross-parameterization via convex-hull domains, we use compatible remeshing algorithms to achieve an accurate approximation of the target geometry and to ensure a complete surface matching. Experimental results show that the compatible meshes constructed are well suited for shape blending and other geometric applications.

In this paper, based on manifold harmonics, we propose a novel framework for 3D shape similarity comparison and partial matching. First, we propose a novel symmetric mean-value representation to robustly construct high-quality manifold harmonic bases on nonuniform-sampling meshes. Then, based on the manifold harmonic bases constructed, a novel shape descriptor is presented to capture both of global and local features of 3D shape. This feature descriptor is isometry-invariant, i.e., invariant to rigid-body transformations and non-rigid bending. After characterizing 3D models with the shape features, we perform 3D retrieval with a up-to-date discriminative kernel. This kernel is a dimension-free approach to quantifying the similarity between two unordered feature-sets, thus especially suitable for our high-dimensional feature data. Experimental results show that our framework can be effectively used for both comprehensive comparison and partial matching among non-rigid 3D shapes.

Small object segmentation is a challenging task in image processing and computer vision. In this paper we propose a visual attention based segmentation approach to segment interesting objects with small size in natural images. Different from traditional methods which use the single feature vectors, visual attention analysis is used on local and global features to extract the region of interesting objects. Within the region selected by visual attention analysis, Gaussian Mixture Model (GMM) is applied to further locate the object region. By incorporation of visual attention analysis into object segmentation, the proposed approach is able to narrow the searching region for object segmentation so as to increase the segmentation accuracy and reduce the computational complex. Experimental results demonstrate that the proposed approach is efficient for object segmentation in natural images, especially for small objects. The proposed method outperforms traditional GMM based segmentation significantly.

Recently, the bag of visual words based image representation is getting popular in object category recognition. Since the codebook of the bag-of-words (BOW) based image representation approach is typically constructed by only measuring the visual similarity of local image features (e.g., k-means), the resulting codebooks may not capture the desired information for object category recognition. This paper proposes a novel optimization method for discriminative codebook construction that considers the category information of local image features as an additional term in traditional visual-similarity-only based codebook construction methods. The category sensitive codebook is constructed through solving an optimization problem. Therefore, the category sensitive codebook construction method goes one step beyond visual-similarity-only methods. Besides, the proposed category sensitive codebook construction method can be implemented with k-means clustering very efficiently and effectively. Experimental results on PASCAL VOC Challenge 2006 data set demonstrate the effectiveness of our method.

In this paper, for every local feature, we propose to learn its similar local features across all positive images, instead of using heuristic distance as similarity measure. Specifically, multiple instance learning (MIL) is employed to simultaneously determine the similar points of a local feature and learn its corresponding discriminative function which can be regarded as some kind of similarity measure. For each local feature, a weak learner is constructed based on such similarity measure. Then AdaBoost selects the most discriminative local features and combines them to form a strong classifier. Experimental results show encouraging performance of our method.

Image classification is an important topic in multimedia analysis, among which multi-label image classification is a very challenging task with respect to the large demand for human annotation of multi-label samples. In this paper, we propose a multi-view multi-label active learning strategy, which integrates the mechanism of active learning and multi-view learning. On one hand we explore the sample and label uncertainties within each view; on the other hand we capture the uncertainty over different views based on multi-view fusion. Then the overall uncertainty along the sample, label and view dimensions are obtained to detect the most informative sample-label pairs. Experimental results demonstrate the effectiveness of the proposed scheme.

With the explosive growth of Web resources, how to mine semantically relevant images efficiently becomes a challenging and necessary task. In this paper, we propose a concept sensitive Markov stationary feature (C-MSF) to represent images and also present a classifier based scheme for web image mining. First, through analyzing the results of Google Image Searcher, we collect an image set, which are highly relevant to a concept. Then the image set is explored to learn a C-MSF about the concept by the algorithm of random walk with restart (RWR), in which the spatial co-occurrence of the bag-of-words representation and the concept information are integrated. Obtaining the concept sensitive representation, SVM is applied to mine the web images, while the highly relevant set are considered as positive examples and other random images as negative ones. Finally, experiments on a crawled web dataset demonstrate the improved performance of the proposed scheme.

In recent years, weblogs (or blogs) have received great popularity worldwide, among which video blogs (or vlogs) are playing an increasingly important role. However, research on vlog analysis is still in the early stage, and how to manage vlogs effectively so that they can be more easily accessible is a challenging problem. In this paper, we propose a novel vlog management model which is comprised of automatic vlog annotation and user-oriented vlog search. For vlog annotation, we extract informative keywords from both the target vlog itself and relevant external resources; besides semantic annotation, we perform sentiment analysis on comments to obtain the overall evaluation. For vlog search, we present saliency-based matching to simulate human perception of similarity, and organize the results by personalized ranking and category-based clustering. An evaluation criterion is also proposed for vlog annotation, which assigns a score to an annotation according to its accuracy and completeness in representing the vlog's semantics. Experimental results demonstrate the effectiveness of the proposed management model for vlogs.

In this paper, we propose a novel contextual descriptor which combines the contextual information and local appearance. Based on Gibbs distribution, a local descriptor is designed. By assembling the contextual information and local descriptors, a new partial contextual descriptor (PCD) is finally presented. Combining Pyramid Match Kernel (PMK) and SVM, we test our new descriptor and obtain higher average precision of classification than using local appearance descriptor.

In recent years, Weblogs (or blogs) have received great popularity worldwide, among which video blogs (or vlogs) are playing an increasingly important role. As vlogs gain in population, how to make them more easily accessible has become a hot research topic. In this paper, we propose a novel automatic annotation model for vlogs. We extract informative keywords from both the target vlog itself and external resources which are semantically and visually relevant to it. We also present a new evaluation criterion, which assigns a score to an annotation according to its accuracy and completeness in representing the vlogpsilas semantics. Experimental results demonstrate the effectiveness of both the annotation model and evaluation criterion.

In this paper, a novel method for object categorization is proposed. We first analyze the phenomenon of large intra-class variation and attribute it to the "subcategory" problem. To reveal the local and distinct properties of the different subcategories, relative spaces are constructed. Then the weighted FLDs (Fisher Linear Discriminant) as weak learners trained in relative spaces are integrated with the boosting framework to form the final classifier. Experiments on 8 categories from Caltech database show the effectiveness of our algorithm.

In most previous works, histograms are simply treated as n-dimensional arrays or even reshaped into vectors when measuring the distances between them. However many histograms have their intrinsic topologies, such as HSV histogram (cone), shape context (polar), orientation histogram (circle). The topologies are important for so-called cross-bin distance, because they determine the similarities between histogram bins, and influence the crossbin distances between histograms. In this paper, we proposed the topologypreserved diffusion distance to take the topology into account. This method extracts the distance by measuring the heat diffusion process defined on the topology of the histogram. Moreover, a fast implementation with time complexity O(N) is developed. Experiments on image retrieval and interest point matching show the effectiveness and efficiency of the proposed method. 1

We propose a novel framework for consistent correspondence between arbitrary manifold meshes. Different from most existing methods, our approach directly maps the connectivity of the source mesh onto the target mesh without needing to segment input meshes, thus effectively avoids dealing with unstable extreme conditions (e.g. complex boundaries or high genus). In this paper, firstly, a novel mean-value Laplacian fitting scheme is proposed, which aims at computing a shape-preserving (conformal) correspondence directly in 3D-to-3D space, efficiently avoiding local optimum caused by the nearest-point search, and achieving good results even with only a few marker points. Secondly, we introduce a vertex relocation and projection approach, which refines the initial fitting result in the way of local conformity. Each vertex of the initial result is gradually projected onto the target model's surface to ensure a complete surface match. Furthermore, we provide a fast and effective approach to automatically detect critic points in the context of consistent correspondence. By fitting these critic points that capture the important features of the target mesh, the output compatible mesh matches the target mesh's profiles quite well. Compared with previous approaches, our scheme is robust, fast, and convenient, thus suitable for common applications.

In relevance feedback, active learning is often used to alleviate the burden of labeling by selecting only the most informative data. Traditional data selection strategies often choose the data closest to the current classification boundary to label, which are in fact not informative enough. In this paper, we propose the moving virtual boundary (MVB) strategy, which is proved to be a more effective way for data selection. The co-SVM algorithm is another powerful method used in relevance feedback. Unfortunately, its basic assumption that each view of the data be sufficient is often untenable in image retrieval. We present our weighted co-SVM as an extension of co-SVM by attaching weight to each view, and thus relax the view sufficiency assumption. The experimental results show that the weighted co-SVM algorithm outperforms co-SVM obviously, especially with the help of MVB data selection strategy.

Image annotation refinement is crucial to improve the performance of automatic image annotation, in which the estimation of word correlation is a key issue. Typically, the word co-occurrence information may be utilized to estimate the word correlation. However, this approach is not accurate enough because it equally treats any word pair co-occurring in the training data and cannot extract synonymy relationship effectively. In this paper, a novel method is developed to estimate the word correlation based on the improved nearest spanning chains (NSC). It can extract more informative and reasonable relations among keywords. Obtaining the enhanced word correlation, a word-based graph is constructed, which is used to re-rank the candidate annotations for an untagged image. Experiments conducted on the typical Corel dataset demonstrate the effectiveness of the proposed method.

Due to the well known semantic gap, content based image retrieval is a difficult problem. To bridge it, relevance feedback as an effective solution has been extensively studied in literatures. However, existing methods follow a single-line searching philosophy, which may lead to a local optimum in search space. To address the problem, we propose a human behavior consistent relevance feedback model for image retrieval in this paper. Simulating human behaviors, the proposed model enable the user to perform relevance feedback in three manners: Follow up, Go back, and Restart. Each manner is a way for the user to provide the system with his or her opinions about search results. The accumulated feedback information can be used to refine the user query and regulate the similarity metric. We adopt the graph ranking algorithm to model the retrieval process. Experiments conducted on standard Corel dataset and Pascal VOC 2006 dataset demonstrate the effectiveness of the proposed mechanism.

Image annotation has been an active research topic in recent years due to its potential impact on both image understanding and web image retrieval. Existing relevance-model-based methods perform image annotation by maximizing the joint probability of images and words, which is calculated by the expectation over training images. However, the semantic gap and the dependence on training data restrict their performance and scalability. In this paper, a dual cross-media relevance model (DCMRM) is proposed for automatic image annotation, which estimates the joint probability by the expectation over words in a pre-defined lexicon. DCMRM involves two kinds of critical relations in image annotation. One is the word-to-image relation and the other is the word-to-word relation. Both relations can be estimated by using search techniques on the web data as well as available training data. Experiments conducted on the Corel dataset and a web image dataset demonstrate the effectiveness of the proposed model.

Most of spectral embedding algorithms such as Isomap, LLE and Laplacian Eigenmap only give map on training samples. One main problem of these methods is to find the embedding of new samples, which is known as the outof-sample problem of spectral embedding. In this paper, we propose a neural network based method to solve this problem. Neural network is used to train and perform both the forward map from high dimensional image space to low dimensional embedding space, and the backward map in the reverse direction. Additionally, combining the forward and backward network, this method is able to build auto-association model to retrieve high dimensional data, and cross association model to learn high dimensional correspondences. Experiments are conducted on real images for forward and backward map, auto-association and cross association.

In this paper, we propose a probabilistic model for Web image mining, which is based on concept-sensitive salient regions without human intervene. Our goal is to achieve a middle-level understanding of image semantics to bridge the semantic gap existing in the field of image mining and retrieval. With the help of a popular search engine, semantically relevant images are collected, and concept-sensitive salient regions are extracted automatically based on an attention model. Then the semantic concept model is learned from the joint distribution of all salient regions with Gaussian mixture model and expectation-maximization algorithm. In addition, by incorporating semantically irrelevant un-salient regions as negative samples, the discriminative power of the solution is further enhanced. Experiments demonstrate the encouraging performance of the proposed method

There are two fundamental problems with the Fisher linear discriminant analysis for face recognition. One is the singularity problem of the within-class scatter matrix due to small training sample size. The other is that it cannot efficiently describe complex nonlinear variations of face images because of its linear property. In this letter, a kernel scatter-difference-based discriminant analysis is proposed to overcome these two problems. We first use the nonlinear kernel trick to map the input data into an implicit feature space F. Then a scatter-difference-based discriminant rule is defined to analyze the data in F. The proposed method can not only produce nonlinear discriminant features but also avoid the singularity problem of the within-class scatter matrix. Extensive experiments show encouraging recognition performance of the new algorithm.

Most face recognition systems focus on photo-based face recognition. In this paper, we present a face recognition system based on face sketches. The proposed system contains two elements: pseudo-sketch synthesis and sketch recognition. The pseudo-sketch generation method is based on local linear preserving of geometry between photo and sketch images, which is inspired by the idea of locally linear embedding. The nonlinear discriminate analysis is used to recognize the probe sketch from the synthesized pseudo-sketches. Experimental results on over 600 photo-sketch pairs show that the performance of the proposed method is encouraging.

Landuse classification is an important problem in the remote sensing field. It can be used in a wide range of applications. In this paper, we propose a hybrid method fusing edges and regions information for the landuse classification of multispectral images. It mainly includes the steps of image pre-processing, initial segmentation and region merging. Especially, a novel spatial mean shift procedure is proposed so that some information can be extracted and used in the successive steps. Aiming at the multispectral images processing, we also design a band weighting strategy that give a proper weight to each band adoptively according to the region to be processed. Experimental results on the Landsat TM and ETM+ images validate the performance of the proposed method.

This paper presents a novel scheme for manifold learning. Different from the previous work reducing data to Euclidean space which cannot handle the looped manifold well, we map the scattered data to its intrinsic parameter manifold by semisupervised learning. Given a set of partially labeled points, the map to a specified parameter manifold is computed by an iterative neighborhood average method called anchor points diffusion procedure (APD). We explore this idea on the most frequently used close formed manifolds, Stiefel manifolds whose special cases include hyper sphere and orthogonal group. The experiments show that APD can recover the underlying intrinsic parameters of points on scattered data manifold successfully.

MOTIVATION: One popular method for analyzing functional connectivity between genes is to cluster genes with similar expression profiles. The most popular metrics measuring the similarity (or dissimilarity) among genes include Pearson's correlation, linear regression coefficient and Euclidean distance. As these metrics only give some constant values, they can only depict a stationary connectivity between genes. However, the functional connectivity between genes usually changes with time. Here, we introduce a novel insight for characterizing the relationship between genes and find out a proper mathematical model, variable parameter regression and Kalman filtering to model it. RESULTS: We applied our algorithm to some simulated data and two pairs of real gene expression data. The changes of connectivity in simulated data are closely identical with the truth and the results of two pairs of gene expression data show that our method has successfully demonstrated the dynamic connectivity between genes. CONTACT: [email protected].

The three-dimensional (3-D) reconstruction of generalized cylinders (GCs) is an important research field in computer vision. One of the main difficulties is that some contour features in images cannot be reconstructed by traditional stereovision because they do not correspond to reflectance discontinuities of surface in space. In this paper, we present a novel, parametric approach for the 3-D reconstruction of circular generalized cylinders (CGCs) only from the limb edges of CGCs in two images. Instead of exploiting the invariant and quasiinvariant properties of some specific subclasses of GCs in projections, our reconstruction is achieved by some general assumptions on GCs, and can, therefore, be applied to a broader subclass of GCs. In order to improve robustness, we perform the extraction and labeling of the limb edge interactively, and estimate the epipolar geometry between two images by an optimal algorithm. Then, for different types of GCs, three kinds of symmetries (parallel symmetry, skew symmetry, and local smooth symmetry) are employed to compute the symmetry of limb edges. The surface points corresponding to limb edges in images are reconstructed by integrating the recovered epipolar geometry and the properties induced from the assumptions that we make on the GCs. Finally, a homography-based method is exploited to further refine the 3-D description of the GC with a coplanar curved axis.

BACKGROUND: Subcellular localization of a new protein sequence is very important and fruitful for understanding its function. As the number of new genomes has dramatically increased over recent years, a reliable and efficient system to predict protein subcellular location is urgently needed. RESULTS: Esub8 was developed to predict protein subcellular localizations for eukaryotic proteins based on amino acid composition. In this research, the proteins are classified into the following eight groups: chloroplast, cytoplasm, extracellular, Golgi apparatus, lysosome, mitochondria, nucleus and peroxisome. We know subcellular localization is a typical classification problem; consequently, a one-against-one (1-v-1) multi-class support vector machine was introduced to construct the classifier. Unlike previous methods, ours considers the order information of protein sequences by a different method. Our method is tested in three subcellular localization predictions for prokaryotic proteins and four subcellular localization predictions for eukaryotic proteins on Reinhardt's dataset. The results are then compared to several other methods. The total prediction accuracies of two tests are both 100% by a self-consistency test, and are 92.9% and 84.14% by the jackknife test, respectively. Esub8 also provides excellent results: the total prediction accuracies are 100% by a self-consistency test and 87% by the jackknife test. CONCLUSIONS: Our method represents a different approach for predicting protein subcellular localization and achieved a satisfactory result; furthermore, we believe Esub8 will be a useful tool for predicting protein subcellular localizations in eukaryotic organisms.

BACKGROUND: Microarray experiments are becoming a powerful tool for clinical diagnosis, as they have the potential to discover gene expression patterns that are characteristic for a particular disease. To date, this problem has received most attention in the context of cancer research, especially in tumor classification. Various feature selection methods and classifier design strategies also have been generally used and compared. However, most published articles on tumor classification have applied a certain technique to a certain dataset, and recently several researchers compared these techniques based on several public datasets. But, it has been verified that differently selected features reflect different aspects of the dataset and some selected features can obtain better solutions on some certain problems. At the same time, faced with a large amount of microarray data with little knowledge, it is difficult to find the intrinsic characteristics using traditional methods. In this paper, we attempt to introduce a combinational feature selection method in conjunction with ensemble neural networks to generally improve the accuracy and robustness of sample classification. RESULTS: We validate our new method on several recent publicly available datasets both with predictive accuracy of testing samples and through cross validation. Compared with the best performance of other current methods, remarkably improved results can be obtained using our new strategy on a wide range of different datasets. CONCLUSIONS: Thus, we conclude that our methods can obtain more information in microarray data to get more accurate classification and also can help to extract the latent marker genes of the diseases for better diagnosis and treatment.

This work is a continuation and extension of our previous research where kernel Fisher discriminant analysis (KFDA), a combination of the kernel trick with Fisher linear discriminant analysis (FLDA), was introduced to represent facial features for face recognition. This work makes three main contributions to further improving the performance of KFDA. First, a new kernel function, called the cosine kernel, is proposed to increase the discriminating capability of the original polynomial kernel function. Second, a geometry-based feature vector selection scheme is adopted to reduce the computational complexity of KFDA. Third, a variant of the nearest feature line classifier is employed to enhance the recognition performance further as it can produce virtual samples to make up for the shortage of training samples. Experiments have been carried out on a mixed database with 125 persons and 970 images and they demonstrate the effectiveness of the improvements.

In this paper, we study the dynamic stereo problem, i.e. to recover the shape of dynamic scene from multiple synchronized image sequences. To incorporate both spatial and temporal information for depth recovery, we propose a statistical framework that uses pixel process model to encode temporal coherence, and Markov random fields (MRFs) for spatial coherence. In this framework, the dynamic depth recovery problem is finally formulated as an optimization problem, and is optimized by using the belief propagation algorithm. Experimental results with the real dynamic scenes illustrate our method's ability of robust shape recovery.

A dike detection method based on active contour model is introduced in this paper for our flood management project. Experimental result on DEM (Digital Elevation Model) of Jingjiang River of China shows that a continuous and closed dike description could be gotten by this method.

This paper introduces an original entropy-based method for water body extraction from multi-source satellite images. The proposed method is remarkably simple and relies only on the assumption that water body in the image is a smooth area. This assumption is most of the time verified in optical and radar images, which makes the approach insensitive to image sensors. Experimental results illustrate the robustness of the proposed approach.

Most current learning algorithms for image retrieval are based on dichotomy relevance judgement (relevant and non-relevant), though this measurement of relevance is too coarse. To better identify the user needs and preference, a good retrieval system should be able to handle multilevel relevance judgement. In this paper, we focus on relevance feedback with multilevel relevance judgment, where the relevance feedback is considered as an ordinal regression problem. Herbrich has proposed a support vector learning algorithm for ordinal regression based on the Linear Utility Model. His algorithm is intrinsically to train a SVM on a new derived training set, whose size increases rapidly when the original training set gets bigger. This property limits its applicability in relevance feedback, due to real-time requirement of the interactive process. By thoroughly analyzing Herbrich's algorithm, we first propose a new model for ordinal regression, called Cascade Linear Utility Model, then a practical SVM-based algorithm for image retrieval upon it. Our new algorithm is tested on a real-world image database, and compared with other three algorithms capable to handle multilevel relevance judgment. The experimental results show that the retrieval performance of our algorithm is comparable with that of Herbrich's algorithm but with only a fraction of its computational time, and apparently outperform the other methods.

Most current relevance feedback algorithms rely only on labeled samples, and ignore the possible role of the distribution of all samples. We investigate the potential role of sample distribution in relevance feedback. We present a probabilistic framework for relevance feedback in content-based image retrieval. Based on Bayes rule, this framework combines the probability densities of relevant samples and those of all samples in a function to rank images. In this way, the characteristic of the distribution of all samples is taken into account to improve the performance. The density estimation is conducted with non-parametric density estimation, and the densities of all samples can be predetermined before any query. The new approach was evaluated on a database of 20,000 images and compared to some current solutions. Experimental results have demonstrated the effectiveness of our approach.

A new method derived from the Frost filter kernel for SAR (synthetic aperture radar) image speckle reduction is introduced. In the speckle reduction procedure, both the statistical information of the central pixel in the filter window and that of the neighborhood pixels are used. The experimental results on ERS-1 PRI images demonstrate the efficiency of the proposed method compared to other common filters.

A simple and efficient method is proposed to visualize a symmetric second order tensor along 3D curves. It is also applied to an oil well model in which, along the well trajectory, the spatial distribution of geostress is visualized to predict the stability of the oil well bore. To give visual representations several forms are used: (1) we visually compare the magnitude of stresses to determine the mud weight; (2) based on a set of ellipse cross-sections along the well trajectory and the ellipticity of the ellipse we predict the shear failure direction and the stability of the oil well bore; (3) from a hinged curved surface, we emphasize the direction of principal stresses; (4) to give useful information we apply other skills such as magnified details, animation and an interactive interface. Our method avoids two defects in the references, that is, the visual cluster caused by discrete point tensor glyphs, and the integration problem of the hyperstreamline.

We propose a general scheme of region competition (GSRC) for image segmentation based on scale space. First, we present a novel classification algorithm to cluster the image feature data according to the generally defined peaks under a certain scale and a scale space-based classification scheme to classify the pixels by grouping the resultant feature data clusters into several classes with a standard classification algorithm. Next, to reduce the resultant segmentation error, we develop a nonparametric probability model from which the functional for GSRC is derived. We also design a general and formal approach to automatically determine the initial regions. Finally, we propose the kernel procedure of GSRC which segments an image by minimizing the functional. The strategy adopted by GSRC is first to label pixels whose corresponding regions can be determined in large likelihood, and then to fine-tune the final regions with the help of the nonparametric probability model, boundary smoothing, and region competition. Although the description of the scheme is nonparametric in this paper, GSRC can also work parametrically if all nonparametric procedures in this paper are substituted with the parametric counterparts.

We present in this paper an approach to extracting a human parametric 2D model for the purpose of estimating human posture and recognizing human activity. This task is done in two steps. In the first step, a human silhouette is extracted from a complex background under a fixed camera through a statistical method. By this method, we can reconstruct the background dynamically and obtain the moving silhouette. In the second step, a genetic algorithm is used to match the silhouette of the human body to a model in parametric shape space. In order to reduce the searching dimension, a layer method is proposed to take the advantage of the human model. Additionally we apply a structure-oriented Kalman filter to estimate the motion of body parts. Therefore the initial population and value in the GA can be well constrained. Experiments on real video sequences show that our method can extract the human model robustly and accurately.

A region based parametric motion estimation approach is proposed. As a feature based approach, sequential images are firstly represented in a multi-scale region hierarchy, then temporally adjacent region sets are matched along scale-space by symbolic optimization of compound criteria based region similarity, so that temporal region correspondence and the parametric motion representation can be obtained. The motion of regions is represented in affine transformation and estimated in a coarse-to-fine manner so that an intrinsic topological constraint can be applied to enhance the stability of the optimization. Finally the regions are grouped to uniform motion region (UMR) by their motion consistency to realize a compact representation. Compared with classical optical flow methods, our approach shows three advantages. First, motion meaning is easy to interpret at high-level. Second, fundamental problems such as aperture problem, boundary problem, or large motion problem can be some what tackled by applying structure information. Third, computation is more stable and efficient because of the symbolic feature-based methodology and the multi-scale framework.

In this paper we show that based on coplanar constraint the motion and structure of the articulated object can be determined within a scale factor from a monocular sequence of images. By genetic algorithm we achieve the unique robust numerical solution to motion estimation of the coplanar links. Then we apply the techniques to human motion analysis, and obtain the 3D motion data of joints, reanimate successfully the data. The experiments with simulated data and real images are included to demonstrate the validity of the theoretic results.

Presents a nonparametric region competition algorithm which combines scale-space clustering and region competition to segment the image. It also proposes a formal and general procedure to automatically find the initial regions. Our algorithm can also segment an image into regions which are not homogeneous in the sense of statistics, but is homogeneous in the sense of semantics with respect to the segmentation context.

A new two-step classification scheme based on nonparametric estimation of density function and scale-space filtering is presented. This scheme is able to combine traditional supervised classification techniques with clustering. After nonparametric estimation of the underlying density function, this scheme utilizes scale-space filtering and a novel classification algorithm to extract the intrinsic basic structure of the data. Then, depending on applications, one of the traditional clustering or classification techniques may be employed to obtain a final high level data structure.

There exists an optical flow in the scale-space if the multiscale representation of an image is viewed as an ordinary image sequence in the time domain. This technique can be used to solve the ill-posed tracking problem in the scale-space.

In this paper we propose a new approach to recovering epipolar geometry from a pair of uncalibrated images. We first detect the feature points. By minimizing a proposed cost function, we match the feature points, discard the outliers and recover the epipolar geometry in one step. Experiments on real images show that this approach is effective and fast.

This paper presents a novel multiscale representation of one-dimensional signal based on complex analysis. We show that a signal and its derivative at different scales can be represented by one analytic function defined on the unit disc in the complex plane, which is called the Schwarz representation of a signal. This representation is applied to the matching problem. Using the theory of analytic functions, we are able to define the inverse of a signal. The matching function between two signals can be defined as the composition of one signal's Schwarz representation and another signal's inverse. The matching function determined by this method has a group structure and is close-formed.

In this paper we propose an evolutionary framework to establish feature correspondence from two uncalibrated images. By minimizing a proposed cost function, we match the feature points, discard the outliers and recover the epipolar geometry in one step. The unifying framework is furnished by genetic algorithm. We also create a new genetic operator to exchange the information between match process and robust epipolar line estimation so that epipolar geometry constraint can be elegantly incorporated into match process. Experiments on synthetic and real images show that this approach is very effective and fast.

An unsupervised model-based image segmentation technique requires the model parameters for the various image classes in an observed image to be estimated directly from the image. The accuracy of the segmentation depends on the correct estimation of the parameters, as well as on the correct labeling of the pixels. In this work, the parameters are estimated by a multiresolution analysis on the histogram and a robust estimator using least median of squares. The labeling process is based on majority game theory. The method is tested in various synthetic and real images, showing its effectiveness.

For the purpose of estimating the posture parameters of a moving human body in visual surveillance applications, we present a model-based shape analysis method and convert the problem to the optimal matching of 2D silhouettes of the human body in parametric shape space. We point out the causality of shape similarity in morphological scale space. Based on this causality and after using the /spl alpha/-/spl beta/-/spl gamma/ filter to estimate the initial value of the matching, the method we proposed works in a coarse-to-fine manner on silhouettes practically segmented from a complex environment, and can converge fast enough to meet real-time requirements.

A new entropy-based thresholding is presented. As a result, a gray level minimizing the difference between the entropy of the object and that of the background will be the desired threshold. Performance of the proposed method has been tested on some ultrasonic images. The results obtained from the method proposed by Kapur et al. (1985) are also given for comparison.

This paper first suggests some general principles for computing perceptual organization and then presents a model to group points into line contours. The grouping factors arise at two levels. At the first level, grouping factors are identified as direct point-point interaction and orientation interaction. Point-point interaction is short-ranged and homogeneous. Orientation interaction is locally oriented and mediated by local visual context. At the second level, grouping factors are global geometric binding effects arising from geometric redundancy reduction. Line perception is then generally formulated as combinatorial optimization. Since it includes local, global interactions and local context effects, the model may capture partially grouping ability of human vision systems.

We propose an efficient Voronoi transform algorithm for constructing Voronoi diagrams using segment lists of rows. A significant feature of the algorithm is that it takes segments rather than pixels as the basic units to represent and propagate the nearest neighbor information. The segment lists are dynamically updated as they are scanned. A distance map can then be easily computed from the segment list representation of the Voronoi diagram. Experimental results have demonstrated its high efficiency. Extension of the algorithm to higher dimensions is also discussed.

This paper's main contributions are three-fold. Firstly, it is shown that the two existing template matching-like definitions of the Hough transform proposed by Princen, Illingworth and Kittler (1992) and by Bergen and Shvaytser (1991) are inadequate. The principal reason behind this is that the common implicit assumption of these two definitions, that every feature point within the template associated with a given accumulator cell E/sub 0/ in Hough space votes equally to E/sub 0/, is not reasonable. Secondly, an inherent probabilistic aspect of the Hough transform embedded in the transformation process from the image space to the parameter space is clarified. It is concluded that when the Hough transform is used to detect a pattern, an appropriate curve (surface, if the number of the parameters to be detected is more than 2) density function, which depends on the parameterization of the pattern, must be implicitly or explicitly provided to eliminate the uncertainties resulting from such a probabilistic aspect. Thirdly, a new framework of the Hough transform is proposed which mainly consists of two parts, namely parameterization and associated curve (surface) density function.

Thin network extraction from three dimensional (3D) images is a new issue in image processing. It is of major importance in medical vascular imaging for diagnostics, therapy planning and surgery. Here, the authors develop a framework for automatic vascular network extraction from the volumic image. The approach consists in treating the 3D image as a hyper-surface of IR/sup 4/. It is shown that the crest points of this hyper-surface correspond to the center line of the thin network in the image. Promising results are shown on synthetic and real vascular images.

We approximate Gaussian function with any scale by linear combination of Gaussian functions with dyadic scales so that scale space can be constructed much more efficiently. The approximation error is so small that our approach can be used widely in computer vision and pattern recognition. Features at any scale can also be found efficiently by tracking from the dyadic scales.

We derive the fundamental constraints of using nonlinear diffusion in early vision. The deformed images from continuous nonlinear diffusion form nonlinear scale space. We first formularize some criteria of this space, and then, to obey these criteria, we derive the constraints on diffusion coefficient. We show that the "positive coefficient", which was considered as the only constraint in most of the literature, is not sufficient because it may introduce spurious edges. We propose diffusing the derivatives of the image, and prove that in these cases, the "positive coefficient" guarantees that no new extreme nor edge will be generated.

In this paper, we propose a novel method for extracting the geometric primitives from geometric data. Specifically, we use tabu search to solve geometric primitive extraction problem. In the best of our knowledge, it is the first attempt that tabu search is used in computer vision. Our tabu search (TS) has a number of advantages: (1) TS avoids entrapment in local minima and continues the search to give a near-optimal final solution; (2) TS is very general and conceptually much simpler than either SA or GA; (3) TS is very easy to implement and the entire procedure only occupies a few lines of code; (4) TS is a flexible framework of a variety of strategies originating from artificial intelligence and is therefore open to further improvement.

3D pose estimation from 2D image data is a fundamental problem in computer vision. In this paper, a pose estimation method based on planar-curved features on the surface of an object is presented. This method is linear and generally applicable to any high degree (/spl ges/2) planar-curved features. So far, the 3D pose estimation methods presented in the literature are based either on point/line features or quadratic-curved features. The methods based on point and line features have to solve the correspondence problem. As there are so many edge points and line segments in an image, to establish the correspondences of these primitives between images is very difficult and time consuming in practice. The methods based on quadratic-curved features have to solve non-linear equations which normally result in many pseudo solutions. To eliminate these extra solutions and choose the right solution is very difficult. Our method is advantageous in these aspects.

We present a method to reconstruct an ellipsoid from its occluding contours observed in three images. We derive a linear relationship between an ellipsoid and its perspective projection. From this relationship, we show that, if cameras are calibrated, an ellipsoid can be reconstructed from its three views and the solution is, in general, unique; if the cameras are weakly calibrated, then the reconstruction is also unique but up to projectivity. Our method has been successfully tested on synthetic data and on real image data.

Thin network extraction from three dimensional images is a new issue in computer vision. It is of major importance in medical vascular imaging for diagnostic, therapy planning and surgery. In this paper, we develop a framework for automatic thin network extraction from the volumic image. The approach consists in treating the 3D image as a hyper-surface of IR/sup 4/. It is shown that the crest points of this hyper-surface correspond to the center line of the thin network in the image. Using a simple mathematical model, we establish the relationship between the computed principal curvatures of the hyper-surface and the geometry of the network. Promising results are shown on synthetic and real vascular images.

A manipulator wrist-mounted camera considerably facilitates motion stereo, object tracking, and active perception. An important issue in active vision is to determine the camera position and orientation relative to the camera platform (head-eye calibration or hand-eye calibration). We present a technique for calibrating the head-eye geometry and the camera intrinsic parameters. The technique allows camera self-calibration because it requires no reference object and directly uses the images of the environment. Camera self-calibration is important especially where the underlying visual tasks do not permit the use of reference objects. Our method exploits the flexibility of the active vision system, and bases camera calibration on a sequence of specially designed motion. It is shown that if the camera intrinsic parameters are known a priori, the orientation of the camera relative to the platform can be solved using 3 pure translational motions. If the intrinsic parameters are unknown, then two sequences of motion, each consisting of three orthogonal translations, are necessary to determine the camera orientation and intrinsic parameters. Once the camera orientation and intrinsic parameters are determined, the position of the camera relative to the platform can be computed from an arbitrary nontranslational motion of the platform. All the computations in our method are linear. Experimental results with real images are presented.

Discusses a nonparametric approach for calibrating a CCD camera, a constrained topological mapping (CTM) approach to analyze the systematic imaging errors of an image system and compare it with parametric approaches which are based on optimization and have been discussed by many other authors. This nonparametric approach has several distinct features. In this approach, some distortion surfaces are derived directly from the training samples. Because no analytical form of these surfaces is assumed, when we modeled the distortions by a nonparametric model, the systematic imaging errors instead of mere lens distortions are considered. This gives an new approach to analyze the imaging errors of a particular imaging system. Experimental results are given in detail, which indicate that both in image projection and in 3D reconstruction, the accuracy is much improved when the nonparametric approach is employed for calibrating a camera.>

In this paper, we proposed a linear-transform-based method to compute 3D moments. Instead of computing the conventional moments directly, we compute the so-called LT moments which are related with the conventional moments by three linear transforms. By choosing appropriately these transforms, it is shown that the computational complexity can be considerably reduced.

In this paper, we present a novel method to estimate the parameters of a quadratic curve. We show that by grouping the data points and by estimating the so-called moments difference (MD) which are unbiased and consistent, the parameters of a quadratic curve can be estimated with higher accuracy than the existing methods.

Fourier descriptor and geometric moments are respectively the contour and region representations of objects in a binary image. In this paper, the authors reveal the relation between these two representations. The authors first generalize the Fourier descriptor to make it suitable for objects with more complex form. Then it is shown that the two representations are related by a simple linear transformation.

We present an algebraic method to reconstruct quadric surfaces from occluding contours observed in two images. The occluding contour is the image of a special curve, called rim, on the surface. It is defined by the fact that the optical rays of their points are tangential to the surface. We show that, although the occluding contours in two images do not correspond to the same rim on the surface, we can reconstruct the surface from its two images by solving three quadratic equations. Our method has been successfully tested by the simulated data and by the real image data.

By implicit camera calibration, we mean the process of calibrating a camera without explicitly computing its physical parameters. Implicit calibration can be used for both three-dimensional (3-D) measurement and generation of image coordinates. In this paper, we present a new implicit model based on the generalized projective mappings between the image plane and two calibration planes. The back-projection and projection processes are modelled separately to ease the computation of distorted image coordinates from known world points. A set of constraints of perspectivity is derived to relate the transformation parameters of the two calibration planes. Under the assumption of the radial distortion model, we present a computationally efficient method for explicitly correcting the distortion of image coordinates in frame buffer without involving the computation of camera position and orientation. By combining with any linear calibration techniques, this method makes explicit the camera physical parameters. Extensive experimental comparison of our methods with the classic photogrammetric method and Tsai's (1986) method in the aspects of 3-D measurement (both absolute and relative errors), the prediction of image coordinates, and the effect of the number of calibration points, is made using real images from 15 different depth values.>

For both 3-D reconstruction and prediction of image coordinates, cameras can be calibrated implicitly without involving their physical parameters. The authors present a two-plane method for such a complete calibration, which models all kinds of lens distortions. First, the modeling is done in a general case without imposing the pinhole constraint. Epipolar curves considering lens distortions are introduced and are found in a closed form. Then, a set of constraints of perspectivity is derived to constrain the modeling process. With these constraints, the camera physical parameters can be related directly to the modeling parameters. Extensive experimental comparisons of the methods with the classic photogrammetric method and Tsai's method relating to the aspects of 3-D measurement, the effect of the number of calibration points, and the prediction of image coordinates, are made using real images from 15 different depth values.>

Fractal Brownian random (FBR) field is an extension of fractional Brownian motion (FBM) and has been successfully used in image analysis, the generation of natural scenes, fractal geometry, and other areas. But its implementation is difficult and limits its applications. In this paper, by extending the AR model, we propose a new approach, the pyramid-AR-model approach, to implement FBR fields. The new method has the same computational complexity as A. Fournier's, but it can generate FBR fields much more accurately.

Presents a stereo vision method based on quadric curves. Different to the previous curve based stereo method in which curves in images are used for computing the features of the points and for establishing the point-point correspondence between the left and the right image, the authors provide a closed form solution for the global matching criterion of two quadric curves in two images and furthermore, a closed form solution for the global reconstruction of the conics in the scene. It is shown that the matching of two quadric curves can be established globally and the reconstruction can be performed by solving three one variable quadric equations. The algorithm can be implemented in real time.>

A unified model for camera calibration is presented. The model does not assume any explicit form of imaging process and can perform linear back projection while considering all lens distortions. This model is well suited to stereo reconstruction. It has been shown that the original two-plane model, the linear and nonlinear pin hole model, can all be realized as special cases of this model. With the two-plane model, optimal orientation of the two reference planes can be found intuitively.>

A method is presented for integrating the problems of matching and motion estimation. A general principle of consistent motion for such a fusion is formulated. Using this principle, the rigidity constraint is incorporated into the correspondence process. A fusion algorithm for point matching is presented. The overdetermination of essential motion equations is utilized directly to give a matching error measurement of any hypothesized correspondence, so that the correspondence and motion estimation are constructed as a single combinatorial optimization problem. A coarse-to-fine strategy for point-matching is also put forward. The method is especially applicable to cases where the set of possible correspondence has been found. Examples show the effectiveness of the proposed method.>

The authors present a unified camera calibration method in which radial distortions can be considered or ignored. The usually tangled process of calibrating intrinsic and extrinsic camera parameters in the zero-distortion case is separated by using vanishing points as calibration points. By ignoring distortions, rough estimates for intrinsic camera parameters are solved linearly. Refinements are made on the basis of the invariance properties of the cross ratio and the collinearity of perspectivity, so that radial distortion can be considered. With the corrected image plane, the usual linear methods can be used to solve extrinsic camera parameters with low computational cost.>

We have previously shown the validity of a statistical model proposed by us for realistic textures. We have also shown how to use this model to map realistic textures onto a still surface of any shape and extent, maintaining consistency and scaling without producing repeating patterns.The problem we have to solve in this paper is how to use this statistical model in order to produce animated objects while keeping temporal consistency. This problem a priori difficult to solve as our mapping technique is statistical.We propose two different techniques. The first method called the image space method creates texture directly on the image planes. The second one called the object space method performs the texture synthesis on the object surface itself which has to be sampled finelly beforehand.

Texture mapping is one of the most important techniques for highquality image synthesis. It can enhance immensely the visual richness of raster-scan images. We address in this paper the problem of mapping planar texture pattern onto arbitrarily curved surfaces. It is well known that such a mapping will inevitably produce distortion for any undevelopable surface. Although many techniques for texture mapping have been presented in the literature [1], to our knowledge, no general method existed which could control or minimize this distortion. We present in this paper a new method for texture mapping. The main idea is to develop the given curved surface approximately on a two-dimensional plane using an approximate locally isometric mapping while minimizing the distortion by an optimization technique. Our experience has shown that by using this technique, one can reduce considerably the distortion and thus improve the quality of the generated images.

This paper presents a new method for the synthesis of textures on 3-D surfaces. With the exception of fractals, graftals and particle systems used for very restricted types of textures, one basic technique for the faithful reproduction of textures has been presented in the literature up to now. In this standard method, textures are synthesized by mapping a rectangular template onto a patch of the curved surface. In the method proposed here, the synthesis is achieved continuously on the surface by interpolating only statistics of the texture which have a much narrower spectrum than the texture itself so that there are no aliasing effects. A vast class of textures can be reproduced with a visual appearance similar to the reference texture used. The method applies for any 3-D surface and is relatively quick and easy.