Search

Abstract: Human hand shape features extraction from image frame sequences is one of the key steps in human hand 2D/3D tracking system and human hand shape recognition system. In order to satisfy the need of human hand tracking in real time, a fast and accurate method for acquirement of edge features from human hand images without consideration of hand over face is put forward in this paper. The proposed approach is composed of two steps, the coarse location phase (CLP) and the refined location phase (RLP) from coarseness to refinement. In the phase of CLP, the hand contour is approximately described by a polygon with concave and convex, an approach to obtaining hand shape polygon using locating points and locating lines is meticulously discussed. Then, a coarse location (CL) algorithm for extraction of interested hand shape features, such as contour, fingertips, roots of fingers, joints and the intersection of knuckle on different fingers, is proposed. In the phase of RLP, a multi-scale approach is introduced into our study to refine the features obtained by the CL algorithm. By means of defining the response strength of different types of features, a refined location (RL) algorithm is proposed. The major contribution of this paper is that the novel detection operators for features of hand images are presented in the above two steps, which have been successfully applied to our 3D hand shape tracking system and 2D hand shape recognition system. A number of comparative studies with real images and online videos demonstrate that the proposed method can extract the three defined human hand image features with high accuracy and high speed. [ABSTRACT FROM AUTHOR]

Abstract: Zernike moments which are superior to geometric moments because of their special properties of image reconstruction and immunity to noise, suffer from several discretization errors. These errors lead to poor quality of reconstructed image and wide variations in the numerical values of the moments. The predominant factor, as observed in this paper, is due to the discrete integer implementation of the steps involved in moment calculation. It is shown in this paper that by modifying the algorithms to include discrete float implementation, the quality of the reconstructed image improves significantly and the first-order moment becomes zero. Low-order Zernike moments have been found to be stable under linear transformations while the high-order moments have large variations. The large variations in high-order moments, however, do not greatly affect the quality of the reconstructed image, implying that they should be ignored when numerical values of moments are used as features. The 11 functions based on geometric moments have also been found to be stable under linear transformations and thus these can be used as features. Pixel level analysis of the images has been carried out to strengthen the results. [Copyright &y& Elsevier]

Recent advances in image classification mostly rely on the use of powerful local features combined with an adapted image representation. Although Convolutional Neural Network (CNN) features learned from ImageNet were shown to be generic and very efficient, they still lack of flexibility to take into account variations in the spatial layout of visual elements. In this paper, we investigate the use of structural representations on top of pretrained CNN features to improve image classification. Images are represented as strings of CNN features. Similarities between such representations are computed using two new edit distance variants adapted to the image classification domain. Our algorithms have been implemented and tested on several challenging datasets, 15Scenes, Caltech101, Pascal VOC 2007 and MIT indoor. The results show that our idea of using structural string representations and distances clearly improves the classification performance over standard approaches based on CNN and SVM with linear kernel, as well as other recognized methods of the literature. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new algorithm for power line identification and extraction from high resolution remote sensing images. Theoretically, it is difficult to detect power lines in satellite images due to some characteristics, such as sub-pixel, weak target, discrete and the complicated background. To our knowledge, the problem of extraction of the power lines from satellite images is faced for the first time. An improved Radon transform, Cluster Radon Transform (CRT), was developed to extract linear feature from satellite image. Compared with conventional Radon transform, CRT can efficiently avoid false alarm. After that, a set of rules of power lines was abstracted to distinguish power lines from other linear feature, such as roads. The experimental results show that CRT not only has strong anti-noise capability to random noise, but also has strong anti-noise capability to system noise caused by non- linear feature. Furthermore, CRT also has the strong capability to detect short segment in an image. Finally, synthetic images and true images were used to verify the new approach. The achievement has potential to be applicable not only to the power line extraction, but also to other weak linear target detection. [ABSTRACT FROM AUTHOR]

Establishing correct correspondences between two faces with different viewpoints has played an important role in 3D face reconstruction and other computer-vision applications. Usually, face images are considered to lack sufficient distinctive features to establish a large number of correspondences on uncalibrated images. In this paper, we investigate pore-scale facial features, which are formed from pores, fine wrinkles, and hair. These features have many characteristics that make them suitable for matching facial images under different variations. Using both biological observation and computer-vision consideration, a new framework is devised for pore-scale facial-feature extraction and matching. The matching difficulty under various skin appearances of different subjects and imaging distortion is also analyzed. For further improving the matching performance and tackling distortions such as varying illuminations and unfocused blurring, a pore-to-pore correspondences dataset is established for training a more distinctive and compact descriptor. Experiments are conducted on a face database containing 105 subjects, and the results prove that the pore-scale features are highly distinctive; face images with a minimum resolution of 600×700 (0.4 mega) pixels contain sufficient details to perform a reliable matching in different poses. Generally, our algorithm can establish between 500 and 2000 correct correspondences on a pair of uncalibrated face images of the same person. Furthermore, the proposed methods can be applied to face recognition, 3D reconstruction, etc. [ABSTRACT FROM AUTHOR]

Abstract: At present, dictionary based models have been widely used in image classification. The image features are approximated as a linear combination of bases selected from the dictionary in a sparse space, resulting in compact patterns. The features applied to image classification usually reside on low dimensional manifolds embedded in a high dimensional ambient space; traditional sparse coding algorithm, however, does not consider this topological structure. It can be characterized naturally by linear coefficients that reconstruct each data point from its neighbors. One of the central issues here is how to determine the neighbors and learn the coefficients. In this paper, the geometrical structures are encoded in two situations. In simple cases when data points distribute on a single manifold, it is explicitly modeled by locally linear embedding algorithm combined with k-nearest neighbors. Nevertheless, in real-world scenarios, complex data points often lie on multiple manifolds. Sparse representation algorithm combined with k-nearest neighbors is instead utilized to construct the topological structures, because it is capable of approximating the data point by selecting its homogenous neighbors adaptively to guarantee the smoothness of each manifold. After obtaining the local fitting relationship, these two topological structures are then embedded into sparse coding algorithm as regularization terms to formulate the corresponding objective functions of dictionary learning on single manifold (DLSM) and dictionary learning on multiple manifolds (DLMM), respectively. Upon this, a coordinate descent scheme is proposed to solve the unified optimization problems. Experimental results on several benchmark data sets, such as Caltech-256, Caltech-101, Scene 15, and UIUC-Sports, show that our proposed algorithms equal or outperform other state-of-the-art image classification algorithms. [Copyright &y& Elsevier]

Abstract: Spatial pyramids have been successfully applied to incorporating spatial information into bag-of-words based image representation. However, a major drawback is that it leads to high dimensional image representations. In this paper, we present a novel framework for obtaining compact pyramid representation. First, we investigate the usage of the divisive information theoretic feature clustering (DITC) algorithm in creating a compact pyramid representation. In many cases this method allows us to reduce the size of a high dimensional pyramid representation up to an order of magnitude with little or no loss in accuracy. Furthermore, comparison to clustering based on agglomerative information bottleneck (AIB) shows that our method obtains superior results at significantly lower computational costs. Moreover, we investigate the optimal combination of multiple features in the context of our compact pyramid representation. Finally, experiments show that the method can obtain state-of-the-art results on several challenging data sets. [Copyright &y& Elsevier]

Abstract: In this paper, we propose a new region-based active contour model (ACM) for image segmentation. In particular, this model utilizes an improved region fitting term to partition the regions of interests in images depending on the local statistics regarding the intensity and the magnitude of gradient in the neighborhood of a contour. By this improved region fitting term, images with noise, intensity non-uniformity, and low-contrast boundaries can be well segmented. Integrated with the duality theory and the anisotropic diffusion process based on structure tensor, a new regularization term is defined through the duality formulation to penalize the length of active contour. By this new regularization term, the structural information of images is utilized to improve the ability of capturing the geometric features such as corners and cusps. From a numerical point of view, we minimize the energy function of our model by an efficient dual algorithm, which avoids the instability and the non-differentiability of traditional numerical solutions, e.g. the gradient descent method. Experiments on medical and natural images demonstrate the advantages of the proposed model over other segmentation models in terms of both efficiency and accuracy. [Copyright &y& Elsevier]

Abstract: Natural image patches are fundamental elements for visual pattern modeling and recognition. By studying the intrinsic manifold structures in the space of image patches, this paper proposes an approach for representing and recognizing objects with a massive number of local image patches (e.g. 17×17 pixels). Given a large collection () of proto image patches extracted from objects, we map them into two types of manifolds with different metrics: explicit manifolds of low dimensions for structural primitives, and implicit manifolds of high dimensions for stochastic textures. We define these manifolds grown from patches as the “”, where corresponds to the perception residual or fluctuation. Using these as features, we present a novel generative learning algorithm by the information projection principle. This algorithm greedily stepwise pursues the object models by selecting sparse and independent (say for each category). During the detection and classification phase, only a small number (say 20) of features are activated by a fast KD-tree indexing technique. The proposed method owns two characters. (1) Automatically generating features () from local image patches rather than designing marginal feature carefully and category-specifically. (2) Unlike the weak classifiers in the boosting models, these selected features are used to explain object in a generative way and are mutually independent. The advantage and performance of our approach is evaluated on several challenging datasets with the task of localizing objects against appearance variance, occlusion and background clutter. [Copyright &y& Elsevier]

Abstract: A novel algorithm that permits the fast and accurate computation of geometric moments on gray-scale images is presented in this paper. The proposed algorithm constitutes an extension of the IBR algorithm, introduced in the past, which was applicable only for binary images. A new image representation scheme, the ISR (intensity slice representation), which represents a gray-scale image as an expansion of several two-level images of different intensity values, enables the partially application of the IBR algorithm to each image component. Moreover, using the resulted set of image blocks, the geometric moments’ computation can be accelerated through appropriate computation schemes. [Copyright &y& Elsevier]

PCA plus LDA is a popular framework for linear discriminant analysis (LDA) in high dimensional and singular case. In this paper, we focus on building a theoretical foundation for this framework. Moreover, we point out the weakness of the previous LDA based methods, and suggest a complete PCA plus LDA algorithm. Experimental results on ORL face image database indicate that the proposed method is more effective than the previous ones. [Copyright &y& Elsevier]