3D Solid Texture Classification Using Locally-Oriented Wavelet Transforms.

Analytical approaches, based on digitised 2D texture models, for an automatic solid (3D) texture synthesis have been recently introduced to Computer Graphics. However, these approaches cannot provide satisfactory solutions in the usual case of natural anisotropic textures (wood grain for example). Indeed, solid texture synthesis requires particular care, and sometimes external knowledge to "guess" the internal structure of solid textures because only 2D texture models are used for analysis. By making some basic assumptions about the internal structure of solid textures, we propose a very efficient method based on a hybrid analysis (spectral and histogram) for an automatic synthesis of solid textures. This new method allows us to obtain high precision solid textures (closely resembling initial models) in a large number of cases, including the difficult case of anisotropic textures.

3D solid textures are important data for computer graphics applications because they are useful for synthesizing photo realistic patterns such as stone, smoke, water, clouds, fire, and wood. Unlike typical 2D textures, 3D solid textures are represented in voxel forms, which use three-dimensional coordinates. Various kinds of research have been conducted for pattern analysis of 2D textures to help development of computer applications. However, pattern analysis of 3D solid textures has not been investigated sufficiently even though 3D data related software is emerging. In this research, we have extended higher order local autocorrelation (HLAC) 2D mask patterns to 3D mask patterns so that pattern analysis software can handle not only 2D texture images but also 3D solid textures. Our experimental system can classify and look for 3D solid textures with similar patterns from a 3D solid texture database by using 3D HLAC masks. We have examined (1) similarity retrieval of 3D solid textures and (2) pattern classification of 3D solid textures by using the system.

Analysis of the Approaches to 3-D Reconstruction of Porous Media

This study investigates local orientation-based approaches to the complex problem of pattern segmentation in three-dimensional (3D) texture image. The current problem focuses on the extraction of so-called lamellar colonies in titanium alloy, which, from the materials science and engineering point of view, are microstructural features that play a fundamental role on crack propagation and bifurcation during mechanical loading. Methods based on local orientation estimation extend the notion of using local gradient to reveal variation of semi-planar pattern orientation in the 3D image. The study introduces a computational approach that accelerates the calculation of the eigenvectors from the local matrices of inertia of all voxels composing the 3D image. Then different paths are proposed to segment colonies or inter-colony boundaries, i.e. polar orientation map and minimum scalar product map, in order to delimitate regions of similar orientations. The investigated segmentation methods have been compared with other methods that are mainly based on the popular solution of filter banks. Tests, which have been performed on both synthetic and real 3D images, show that the proposed local orientation-based methods better delineate object boundaries than the counterparts.

In this paper, Gabor filtering and linear local tangent space alignment algorithm and its improved algorithm are used on face recognition. The Gabor wavelet transform can detect the image information in different directions and scales, according to its selective direction and frequency characteristics. The LLTSA reduces the dimension of the sample while the LLTSA and the other improved algorithms extract secondary feature. Experiment and analyze the average recognition rate of the LLTSA and its improved algorithms with the variation of dimension. The experiment results show the effectiveness of the method, increasing the face recognition accuracy.

Local dominant orientation estimation is one of the most important operations in almost all automatic fingerprint authentication systems. Robust orientation and anisotropy estimation improves the system's reliability in handling low-quality fingerprints, which is crucial for the system's massive application such as securing multimedia. This paper analyzes the robustness of the orientation and anisotropy estimation methods and the effect of the modulus normalization on the estimation performance. A two-stage averaging framework with block-wise modulus handling is introduced to inherit the merits of the both linear and normalized averaging methods. We further propose to set the modulus of an orientation vector to be its anisotropy estimate instead of unity so that the orientation inconsistency of gradients is included in the second stage of averaging. These two measures improve the robustness of the fingerprint local dominant orientation estimation and lead to an anisotropy estimate that reflects the characteristics of fingerprint more effectively. In addition, the proposed approach is computationally efficient for online fingerprint authentication. Extensive experiments using both synthetic images and real fingerprints verify the feasibility of the proposed approach and demonstrate its robustness to noise and low-quality fingerprints.

A unified differential geometric framework for estimation of local surface shape and orientation from projective distortion is proposed, based on a differential version of the stationarity assumption introduced by Malik and Rosenholtz. This framework allows the information content of the gradient of any descriptor defined in a local coordinate frame to be characterized in a very compact form. The analysis encompasses both full affine descriptors and the classical texture gradients. For estimation of local surface orientation and curvature from uncertain observations of affine distortion, the proposed framework allows the dimensionality of the search space to be reduced from five to one. >

Biomedical image analysis has become critically important to the public health and welfare. However, analyzing biomedical images is time-consuming and labor-intensive, and has long been performed manually by highly trained human experts. As a result, there has been an increasing interest in applying machine learning to automate biomedical image analysis. Recent progress in deep learning research has catalyzed the development of machine learning in learning discriminative features from data with minimum human intervention. Many deep learning models have been designed and achieved superior performance in various data analysis applications. This chapter starts with the basic of deep learning models and some practical strategies for handling biomedical image applications with limited data. After that, case studies of deep feature extraction for gene expression pattern image annotations, imaging data completion for brain disease diagnosis, and segmentation of infant brain tissue images are discussed to demonstrate the effectiveness of deep learning in biomedical image analysis.

Compared with 2D texture method, solid texturing can achieve better realistic results. A novel solid texture synthesis method based on 2D exemplar and procedural noise is presented. First, the movement path for exemplar in 3D space is designed using procedural noise. Next, all the pixels of exemplar are used to color the 3D space through their trajectory so that the solid texture is synthesized. Experimental results show that our method is able to efficiently synthesize 3D texture to simulate the natural properties of the given 2D exemplar. The method is simple and stable, and takes on a few seconds on an ordinary PC.

In this paper, we present an improved synthesis algorithm for both 2D and solid textures. Based on the wavelet-based multi-resolution pyramids and the improved optimization of texture synthesis, a coarse texture was synthesized as a skeleton. By adding detail information to the coarse texture, the higher resolution results can be generated. This skeletal growth-like procedure makes the result more controllable. In addition, to make full use of the spatial relationships among texture pixels, wavelet coefficients rather than traditional RGB channels are used to search the nearest neighbor pixels. Experimental study shows that the method plays well for both 2D and solid texture synthesis.

We describe the use of traditional stereological methods to synthesize 3D solid textures from 2D images of existing materials. We first illustrate our approach for aggregate materials of spherical particles, and then extend the technique to apply to particles of arbitrary shapes. We demonstrate the effectiveness of the approach with side-by-side comparisons of a real material and a synthetic model with its appearance parameters derived from its physical counterpart. Unlike ad hoc methods for texture synthesis, stereology provides a disciplined, systematic basis for predicting material structure with well-defined assumptions.

Intense interest in applying convolutional neural networks (CNNs) in biomedical image analysis is wide spread, but its success is impeded by the lack of large annotated datasets in biomedical imaging. Annotating biomedical images is not only tedious and time consuming, but also demanding of costly, specialty-oriented knowledge and skills, which are not easily accessible. To dramatically reduce annotation cost, this paper presents a novel method called AIFT (active, incremental fine-tuning) to naturally integrate active learning and transfer learning into a single framework. AIFT starts directly with a pre-trained CNN to seek "worthy" samples from the unannotated for annotation, and the (fine-tuned) CNN is further fine-tuned continuously by incorporating newly annotated samples in each iteration to enhance the CNN's performance incrementally. We have evaluated our method in three different biomedical imaging applications, demonstrating that the cost of annotation can be cut by at least half. This performance is attributed to the several advantages derived from the advanced active and incremental capability of our AIFT method.

Image denoising represents a crucial initial step in biomedical image processing and analysis. Denoising belongs to the family of image enhancement methods (Bovik, 2009) which comprise also blur reduction, resolution enhancement, artefacts suppression, and edge enhancement. The motivation for enhancing the biomedical image quality is twofold. First, improving the visual quality may yield more accurate medical diagnostics, and second, analytical methods, such as segmentation and content recognition, require image preprocessing on the input. Gradually, noise reduction methods developed in other research fields find their usage in biomedical applications. However, biomedical images, such as images obtained from computed tomography (CT) scanners, are quite specific. Modelling noise based on the first principles of image acquisition and transmission is a too complex task (Borsdorf et al., 2009), and moreover, the noise component characteristics depend on the measurement conditions (Bovik, 2009). Additionally, noise reduction must be carried out with extreme care to avoid suppression of the important image content. For this reason, the results of biomedical image denoising should be consulted with medical experts.

This paper proposes new estimators for the local orientation of elongated structures. Classical methods derive local orientation from the argument of gradient estimator. As they were originally designed for edge detection, such operators can lead to high bias in orientation estimation. We introduce the line orientation operator which applies an oriented model made with straight parallel lines on the local neighborhood. We present results either on synthetic and real images. We compare this new operator with other classical orientation operators.

Multiobjective optimization methods in image analysis are one of the active research domains in the current years. These methods are used for the decision-making process in case of image segmentation. Multiobjective techniques are popular and suitable model for many difficult optimization problems. In various practical problems, different objectives are to be considered. Now, most of the problems have some objectives those are conflicting in nature. Hence, only one objective cannot be optimized or prioritize because it can result in some adverse effect on other objective, and can produce some undesired results in terms of the other objectives. The main goal of this chapter is to give a comprehensive study of multiobjective optimization techniques in biomedical image analysis problem. The different models are categorized depending on the relevant features. For example, the different aspects of the optimization methods employed, different formulations of the problems, categories of data, and the domain of the application. This study mainly focuses on the multiobjective optimization techniques that can be used to analyze digital images specially biomedical images. Here, some of the problems, and challenges related to images are diagnosed and analyzed with multiple objectives. It is a comprehensive study that consolidated some of the recent works along with future directions.