Presented is a novel approach to automatically generate intermediate image descriptors by exploiting concept co-occurrence patterns in the pre-labeled training set that renders it possible to depict complex scene images semantically. This work is motivated by the fact that multiple concepts that frequently co-occur across images form patterns which could provide contextual cues for individual concept inference. The co-occurrence patterns were discovered as hierarchical communities by graph modularity maximization in a network with nodes and edges representing concepts and co-occurrence relationships separately. A random walk process working on the inferred concept probabilities with the discovered co-occurrence patterns was applied to acquire the refined concept signature representation. Through experiments in automatic image annotation and semantic image retrieval on several challenging datasets, the effectiveness of the proposed concept co-occurrence patterns as well as the concept signature representation in comparison with state-of-the-art approaches was demonstrated.

Described is the development of an automated moth species identification and retrieval system (SPIR) using computer vision and pattern recognition techniques. The core of the system was a probabilistic model that infered Semantically Related Visual (SRV) attributes from low-level visual features of moth images in the training set, where moth wings were segmented into information-rich patches from which the local features were extracted, and the SRV attributes were provided by human experts as ground-truth. For the large amount of unlabeled test images in the database or added into the database later on, an automated identification process was evoked to translate the detected salient regions of low-level visual features on the moth wings into meaningful semantic SRV attributes. We further proposed a novel network analysis based approach to explore and utilize the co-occurrence patterns of SRV attributes as contextual cues to improve individual attribute detection accuracy. Working with a small set of labeled training images, the approach constructed a network with nodes representing the SRV attributes and weighted edges denoting the co-occurrence correlation.

Searching large databases using local binary patterns for face recognition has been problematic due to the cost of the linear search, and the inadequate performance of existing indexing methods. We present Discrete Cosine Transform (DCT) hashing for creating index structures for face descriptors. Hashes play the role of keywords: an index is created, and queried to find the images most similar to the query image. It is shown in this research that DCT hashing has significantly better retrieval accuracy and it is more efficient compared to other popular state-of-the-art hash algorithms.

A new automated identification and retrieval system is proposed that aims to provide entomologists, who manage insect specimen images, with fast computer-based processing and analyzing techniques. Several relevant image attributes were designed, such as the so-called semantically-related visual (SRV) attributes detected from the insect wings and the co-occurrence patterns of the SRV attributes which are uncovered from manually labeled training samples. A joint probabilistic model is used as SRV attribute detector working on image visual contents. The identification and retrieval of moth species are conducted by comparing the similarity of SRV attributes and their co-occurrence patterns. The prototype system used moth images while it can be generalized to any insect species with wing structures. The system performed with good stability and the accuracy reached 85% for species identification and 71% for content-based image retrieval on a entomology database.

In recent years, extensive efforts have been made for face recognition and retrieval systems. However, there remain several challenging tasks for face image retrieval in unconstrained databases where the face images were captured with varying poses, lighting conditions, etc. In addition, the databases are often large-scale, which demand efficient retrieval algorithms that have the merit of scalability. To improve the retrieval accuracy of the face images with different poses and imaging characteristics, we introduce a novel feature extraction method to bag-of-words (BoW) based face image retrieval system. It employs various scales of features simultaneously to encode different texture information and emphasizes image patches that are more discriminative as parts of the face. Moreover, the overlapping image patches at different scales compensate for the pose variation and face misalignment. Experiments conducted on a large-scale public face database demonstrate the superior performance of the proposed approach compared to the state-of-the-art method.

We propose an integrated and personalized video retrieval and summarization system. We estimate and impose appropriate preference values on affinity propagation graph of the video frames. Then, our system produces the summary which is useful for the user in her/his relevance feedback and for the retrieval module for comparing video pairs. The experiments confirm the effectiveness of our approach for various query types.

We introduce a novel approach that allows the retrieval of complex images by integrating visual and semantic concepts. The basic idea consists of three aspects. First, we measure the relatedness of semantic and visual concepts and select the visually separable semantic concepts as elements in the proposed image signature representation. Second, we demonstrate the existence of concept co-occurrence patterns. We propose to uncover those underlying patterns by detecting the communities in a network structure. Third, we leverage the visual and semantic correspondence and the co-occurrence patterns to improve the accuracy and efficiency for image retrieval. We perform experiments on two popular datasets that confirm the effectiveness of our approach.

We addresses the problem of concept learning for semantic image retrieval in this research. Two types of semantic concepts are introduced in our system: the individual concept and the scene concept. The individual concepts are explicitly provided in a vocabulary of semantic words, which are the labels or annotations in an image database. Scene concepts are higher level concepts which are defined as potential patterns of cooccurrence of individual concepts. Scene concepts exist since some of the individual concepts co-occur frequently across different images. This is similar to human learning where understanding of simpler ideas is generally useful prior to developing more sophisticated ones. Scene concepts can have more discriminative power compared to individual concepts but methods are needed to find them. A novel method for deriving scene concepts is presented. It is based on a weighted concept co-occurrence network (graph) with detected community structure property. An image similarity comparison and retrieval framework is described with the proposed individual and scene concept signature as the image semantic descriptors. Extensive experiments are conducted on a publicly available dataset to demonstrate the effectiveness of our concept learning and semantic image retrieval framework.

In content-based image retrieval, precision is usually regarded as the top metric used for performance measurement. With image databases reaching hundreds of millions of records, it is apparent that many retrieval strategies will not scale. Data representation and organization has to be better understood. This research focuses on: (a) feature selection and optimal representation of features and (b) multidimensional tree indexing structure. The paper proposes the use of a forward and conditional backward searching feature selection algorithm. The data is then put through a minimum description length based optimal non-uniform bit allocation algorithm to reduce the size of the stored data, while preserving the structure of the data. The results of our experiments show that the proposed feature selection process with a minimum description length based non-uniform bit allocation method gives a system that improves retrieval time and precision.

We propose a new content based image retrieval (CBIR) system combined with relevance feedback and the online feature selection procedures. A measure of inconsistency from relevance feedback is explicitly used as a new semantic criterion to guide the feature selection. By integrating the user feedback information, the feature selection is able to bridge the gap between low-level visual features and high-level semantic information, leading to the improved image retrieval accuracy. Experimental results show that the proposed method obtains higher retrieval accuracy than a commonly used approach.

Expectation-Maximization (EM) algorithms have several limitations, including the curse of dimensionality and the convergence at a local maximum. In this article, we propose a novel learning approach, namely Coevolutionary Feature Synthesized Expectation-Maximization (CFS-EM), to address the above problems. Experiments on real image databases show that CFS-EM out performs Radial Basis Function Support Vector Machine (RBF-SVM), CGP, Discriminant-EM (D-EM) and Transductive-SVM (TSVM) in the sense of classification performance and it is computationally more efficient than RBF-SVM in the query phase.

In this paper, we propose a unified framework of a novel learning approach, namely Coevolutionary Feature Synthesized Expectation-Maximization (CFSEM), to achieve satisfactory learning in spite of these difficulties. The CFS-EM is a hybrid of coevolutionary genetic programming (CGP) and EM algorithm. The advantages of CFS-EM are: 1) it synthesizes low-dimensional features based on CGP algorithm, which yields near optimal nonlinear transformation and classification precision comparable to kernel methods such as the support vector machine (SVM); 2) the explicitness of feature transformation is especially suitable for image retrieval because the images can be searched in the synthesized low-dimensional space, while kernel-based methods have to make classification computation in the original high-dimensional space; 3) the unlabeled data can be boosted with the help of the class distribution learning using CGP feature synthesis approach. Experimental results show that CFS-EM outperforms pure EM and CGP alone, and is comparable to SVM in the sense of classification. It is computationally more efficient than SVM in query phase. Moreover, it has a high likelihood that it will jump out of a local maximum to provide near optimal results and a better estimation of parameters.

We propose an image relevance reinforcement learning (IRRL) model for integrating existing RF techniques in a content-based image retrieval system. Various integration schemes are presented and a long-term shared memory is used to exploit the retrieval experience from multiple users. The experimental results manifest that the integration of multiple RF approaches gives better retrieval performance than using one RF technique alone. Further, the storage demand is significantly reduced by the concept digesting technique.

In this paper, we investigate the effectiveness of coevolutionary genetic programming (CGP) in synthesizing feature vectors for image databases from traditional features that are commonly used. The transformation for feature dimensionality reduction by CGP has two unique characteristics for image retrieval: 1) nonlinearlity: CGP does not assume any class distribution in the original visual feature space; 2) explicitness: unlike kernel trick, CGP yields explicit transformation for dimensionality reduction so that the images can be searched in the low-dimensional feature space. The experimental results on multiple databases show that (a) CGP approach has distinct advantage over the linear transformation approach of Multiple Discriminant Analysis (MDA) in the sense of the discrimination ability of the low-dimensional features, and (b) the classification performance using the features synthesized by our CGP approach is comparable to or even superior to that of support vector machine (SVM) approach using the original visual features.

Managing and manipulating uncertainty in spatial databases are important problems for various practical applications of geographic information systems. We present a probability-based method to model and index uncertain spatial data. In this scheme, each object is represented by a probability density function (PDF) and a general measure is proposed for measuring similarity between the objects. To index objects, an optimized Gaussian mixture hierarchy (OGMH) is designed to support both certain/uncertain data and certain/uncertain queries. As an example of uncertain query support OGMH is applied to the Mojave Desert endangered species protection real dataset. It is found that OGMH provides more selective, efficient and flexible search than the results provided by the existing trial and error approach for endangered species habitat search.

Managing and manipulating uncertainty in spatial databases are important problems for various practical applications. Unlike the traditional fuzzy approaches in relational databases, in this research we propose a probability-based method to model and index uncertain spatial data where every object is represented by a probability density function (PDF). To index PDFs, we construct an optimized Gaussian mixture hierarchy (OGMH) and two variants of uncertain R-tree. We provide a comprehensive comparison among these three indices and plain R-tree on TIGER/Line Southern California landmark point dataset. We find that uncertain R-tree is the best for fixed query and OGMH is suitable for both certain and uncertain queries. Moreover, OGMH is suitable not only for spatial databases, but also for multi-dimensional indexing applications like content based image retrieval, where R-tree is inefficient in high dimensions.

A challenging topic in content-based image retrieval is to determine the discriminant features that improve classification performance. An approach to learn concepts is by estimating mixture model for image databases using EM algorithm; however, this approach is impractical to be implemented for large databases due to the high dimensionality of the feature space. Based on the over-splitting nature of our EM algorithm and the Bayesian analysis of the multiple users' labelling information derived from their relevance feedbacks, we propose a probabilistic MDA to find the discriminating features, and integrate it with the EM framework. The experimental results on Corel images show the effectiveness of concept learning with the probabilistic MDA, and the improvement of the retrieval performance.

Concept learning in content-based image retrieval (CBIR) systems is a challenging task. This paper presents an active concept learning approach based on mixture model to deal with the two basic aspects of a database system: changing (image insertion or removal) nature of a database and user queries. To achieve concept learning, we develop a novel model selection method based on Bayesian analysis that evaluates the consistency of hypothesized models with the available information. The analysis of exploitation vs. exploration in the search space helps to find optimal model efficiently. Experimental results on Corel database show the efficacy of our approach.

Spatial databases typically assume that the positional attributes of spatial objects are precisely known. In practice, however, they are known only approximately, with the error depending on the nature of the measurement and the source of data. In this paper, we address the problem how to perform spatial database operations in the presence of uncertainty. We first discuss a probabilistic spatial data model to represent the positional uncertainty. We then present a method for performing the probabilistic spatial join operations, which, given two uncertain data sets, find all pairs of polygons whose probability of overlap is larger than a given threshold. This method uses an R-tree based probabilistic index structure (PrR-tree) to support probabilistic filtering, and an efficient algorithm to compute the intersection probability between two uncertain polygons for the refinement step. Our experiments show that our method achieves higher accuracy than methods based on traditional spatial joins, while reducing overall cost by a factor of more than two.

The task of a content-based image retrieval (CBIR) system is to cater to users who ezpect to get relevant images with high precision and eficiency in response to query images. This paper presents a concept learning approach that integrates a mixture model of the data, relevance feedback and long-term continuous learning. The concepts are incrementally refined with increased retrieval ezperiences. The concept knowledge can be immediately tmnsplanted to deal with the dynamic database situations such as insertion of new images, removal of ensting images and query images which are outside the database. Experimental results on Core1 database show the eficacy of our approach.

One of the main tasks in content-based image retrieval (CBIR) is to reduce the gap between low-level visual features and high-level human concepts. This research presents a new semi-supervised EM algorithm (NSSEM), where the image distribution in feature space is modeled as a mixture of Gaussian densities. Due to the statistical mechanism of accumulating and processing meta knowledge, the NSS-EM algorithm with longterm learning of mixture model parameters can deal with the cases where users may mislabel images during relevance feedback. Our approach that integrates mixture model of the data, relevance feedback and longterm learning helps to improve retrieval performance. The concept learning is incrementally refined with increased retrieval experiences. Experiment results on Corel database show the efficacy of our proposed concept learning approach.

The paper proposes a content-based image retrieval system which can learn visual concepts and refine them incrementally with increased retrieval experiences captured over time. The approach consists of using fuzzy clustering for learning concepts in conjunction with statistical learning for computing "relevance" weights of features used to represent images in the database. As the clusters become relatively stable and correspond to human concept distribution, the system can yield fast retrievals with higher precision. The paper presents a discussion on problems such as the system mistakenly indentifying a concept, a large number of trials to achieve clustering, etc. Experiments on synthetic data and real image database demonstrate the efficacy of this approach.

This paper presents a model-based approach, which efficiently retrieves correct hypotheses using properties of triangles formed by the triplets of minutiae as the basic representation unit. We show that the uncertainty of minutiae locations associated with feature extraction and shear does not affect the angles of a triangle arbitrarily. Geometric constraints based on characteristics of minutiae are used to eliminate erroneous correspondences. We present an analysis to characterize the discriminating power of our indexing approach. Experimental results on fingerprint images of varying quality show that our approach efficiently narrows down the number of candidate hypotheses in the presence of translation, rotation, scale, shear, occlusion and clutter.

Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback is used to locally estimate the strength of features along each dimension while taking into consideration the correlation between features. In addition to exploring and exploiting local principal information, the system seeks a global space for efficient independent feature analysis by combining such local information.

Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback, in the form of accept or reject examples generated in response to a query image, is used to locally estimate the strength of features along each dimension. This results in local neighborhoods that are constricted along feature dimensions that are most relevant, while enlongated along less relevant ones. We provide experimental results that demonstrate the efficacy of our technique using real-world data.

Most of the current image retrieval systems use “one-shot” queries to a database to retrieve similar images. Typically a K-nearest neighbor kind of algorithm is used, where weights measuring feature importance along each input dimension remain fixed (or manually tweaked by the user), in the computation of a given similarity metric. In this paper, we present a novel probabilistic method that enables image retrieval procedures to automatically capture feature relevance based on user's feedback and that is highly adaptive to query locations. Experimental results are presented that demonstrate the efficacy of our technique using both simulated and real-world data.