Learning texture discrimination rules in a multiresolution system

We describe a texture analysis system in which informative discrimination rules are learned from a multiresolution representation of time textured input. The system incorporates unsupervised and supervised learning via statistical machine learning and rule-based neural networks, respectively. The textured input is represented in the frequency-orientation space via a log-Gabor pyramidal decomposition. In the unsupervised learning stage a statistical clustering scheme is used for the quantization of the feature-vector attributes. A supervised stage follows in which labeling of the textured map is achieved using a rule-based network. Simulation results for the texture classification task are given. An application of the system to real-world problems is demonstrated     

Learning of somatosensory representations for texture discrimination using a temporal coherence principle

In order to perform appropriate actions, animals need to quickly and reliably classify their sensory input. How can representations suitable for classification be acquired from statistical properties of the animal's natural environment? Akin to behavioural studies in rats, we investigate this question using texture discrimination by the vibrissae system as a model. To account for the rat's active sensing behaviour, we record whisker movements in a hardware model. Based on these signals, we determine the response of primary neurons, modelled as spatio-temporal filters. Using their output, we train a second layer of neurons to optimise a temporal coherence objective function. The performance in classifying textures using a single cell strongly correlates with the cell's temporal coherence; hence output cells outperform primary cells. Using a simple, unsupervised classifier, the performance on the output cell population is same as if using a sophisticated supervised classifier on the primary cells. Our results demonstrate that the optimisation of temporal coherence yields a representation that facilitates subsequent classification by selectively conveying relevant information.     

A random walk procedure for texture discrimination

We consider the problem of texture discrimination. Random walks are performed in a plain domain D bounded by an absorbing boundary ? and the absorption distribution is calculated. Measurements derived from such distributions are the features used for discrimination. Both problems of texture discrimination and edge segment detection can be solved using the same random walk approach. The border distributions and their differences with respect to a homogeneous image can classify two different images as having similar or dissimilar textures. The existence of an edge segment is concluded if the boundary distribution for a given window (subimage) differs significantly from the boundary distribution for a homogeneous (uniform grey level) window. The random walk procedure has been implemented and results of texture discrimination are shown. A comparison is made between results obtained using the random walk approach and the first-or second-order statistics, respectively. The random walk procedure is intended mainly for the texture discrimination problem, and its possible application to the edge detection problem (as shown in this paper) is just a by-product.     

Profile-based roughness discrimination with pen-type texture sensor

Tactile discriminations of surface roughness using artificial sensors have been challenging. The modeling methods and parameters that have been using to describe the mechanical properties of rough surface are insufficient for haptic roughness. This paper proposes a method to characterize surface roughness based on the profiles of the surface. A compact handheld pen-type texture sensor with a right probe is developed for the measurement of surface profiles. Based on the contact force and the motion of the senor, profiles in the paths of scanning are estimated. The height variations of a profile are converted to a series of tactile stimuli to represent the contact stimulations in haptic explorations. The mean and the standard deviation of the amplitudes of stimuli are identified as haptic features that indicate the required tangential force to slide on the rough surface and how rough the surface is, respectively. Experiments show that the roughness on four kinds of sandpapers can be clearly distinguished by the proposed discrimination method.     

An evolutionary patch pattern approach for texture discrimination

A new evolutionary approach is presented, based on implicit pattern–process relationships. For implementing this approach, any gray level texture image is decomposed into a progressive sequence of binary patch patterns that describe a process of change from background to foreground domination. Each of the binary patterns throughout these sequences is parameterized, using several metrics that describe, for example, its fragmentation level, both for the background (e.g., white) and foreground (e.g., black) patch patterns. Any texture type is then assumed to have a unique evolutionary path represented by a distinctive region in the feature space of metrics characterizing these patterns and their change. Application of hierarchical clustering based on a few (3 or 4) metrics representing characteristic stages in the patterns’ change process allowed us to accurately discriminate between 50 samples of 10 Brodatz texture types.     

Texture models and image measures for texture discrimination

The task of texture segmentation is to identify image curves that separate different textures. To segment textured images, one must first be able to discriminate textures. A segmentation algorithm performs texture-discrimination tests at densely spaced image positions, then interprets the results to localize edges. This article focuses on the first stage, texture discrimination.We distinguish between perceptual and physical texture differences: the former differences are those perceived by humans, while the latter, on which we concentrate, are those defined by differences in the processes that create the texture in the scene. Physical texture discrimination requires computing image texture measures that allow the inference of physical differences in texture processes, which in turn requires modeling texture in the scene. We use a simple texture model that describes textures by distributions of shape, position, and color of substructures. From this model, a set of image texture measures is derived that allows reliable texture discrimination. These measures are distributions of overall substructure length, width, and orientation; edge length and orientation; and differences in averaged color. Distributions are estimated without explicitly isolating image substructures. Tests of statistical significance are used to compare texture measures.A forced-choice method for evaluating texture measures is described. The proposed measures provide empirical discrimination accuracy of 84 to 100% on a large set of natural textures. By comparison, Laws' texture measures provide less than 50% accuracy when used with the same texture-edge detector. Finally, the measures can distinguish textures differing in second-order statistics, although those statistics are not explicitly measured.The author was with the Robotics Laboratory, Computer Science Department, Stanford University, Stanford, California 94305. He is now with the Institut National de Recherche en Informatique et en Automatique (INRIA), Sophia-Antipolis, 2004 Route des Lucioles, 06565 Valbonne Cedex, France.     

Texture discrimination based upon an assumed stochastic texture model

A new approach to texture discrimination is described. This approach is based upon an assumed stochastic model for texture in imagery and is an approximation to the statistically optimum maximum likelihood classifier. The construction and properties of the stochastic texture model are described and a digital filtering implementation of the resulting maximum likelihood texture discriminant is provided. The efficacy of this approach is demonstrated through experimental results obtained with simulated texture data. A comparison is provided with more conventional texture discriminants under identical conditions. The implications to texture discrimination in realworld imagery are discussed.     

基于纹理特征的自然图像鉴别方法*

采用图像的纹理特性建立模型,实现了自然图像和人工图像（主要是简笔画、表格、函数图像）的鉴别。首先,计算两类图像的纹理特征,统计并分析数据找出规律;其次,选择图像的平滑度、一致性和图像熵作为代表性的纹理特征,组建数学模型;最后,利用三个特征的线性组合构成鉴别函数,并用模拟退火算法确定权值。实验表明,该方法可以很好地鉴别两类图像,识别效率较高。     

Spectral texture for improved class discrimination in complex terrain

Abstract

A spatial co-occurrence algorithm has been used to derive image texture from Landsat Multispectral Scanner (MSS) data to increase classification accuracy in a moderate relief, boreal environment in eastern Canada. The aim was to investigate ‘data-driven improvements’, including those available through digital elevation modelling. Overall classification accuracy using MSS data alone was 59·1 per cent when compared to a biophysical inventory of the area compiled primarily by aerial photointerpretation. This increased to 66·2 per cent with MSS plus texture and to 89·8 per cent when MSS data were analysed with geomorphometry extracted from a digital elevation model (DEM). The introduction of MSS texture resulted in statistically significant increases in individual class accuracies in classes that were also well defined using the geomorphometric and integrated data sets. This suggested that some of the additional information provided by geomorphometry was also contained in spectral texture. It was also noted that individual texture orientations resulted in higher class accuracies than average texture measures; this is probably related to structural (slope/aspect) characteristics of specific vegetation communities.     

Penumbra masks

Computation of physically-based shadows can be significantly accelerated by limiting computations into regions where penumbras appear. In this paper, we present a general penumbra detection method that efficiently bounds regions where penumbras occur in a shared projection plane of an area light source. We introduce a novel area filling operator, which allows effective and conservative area masking with respect to all viewpoints, i.e., sampling points within a planar polygonal light source. The area filling operator uses a point sprite rendering technique on a set of silhouette boundaries to create a penumbra mask, which is essentially a modified occlusion map. We show how to efficiently test the geometry and screen-space pixels against the penumbra mask. An important advantage of our method is that we can separate lit and umbra regions, and thus drive various soft shadow algorithms to focus their computational efforts into potential penumbras. Due to the relative simplicity of computations, penumbra masks can be efficiently generated with graphics hardware. As an example, we accelerate shadow map supersampling to demonstrate significant speedups that utilizations of penumbra masks provide.     

基于纹理信息的室内场景语义标注学习方法

针对目前室内场景视频中关键物体的检测、跟踪及信息编辑等方面主要是采用人工处理方式,存在效率低、精度不高等问题,提出了一种基于纹理信息的室内场景语义标注学习方法。首先,采用光流方法获取视频帧间的运动信息,利用关键帧标注和帧间运动信息进行非关键帧的标注初始化;然后,利用非关键帧的图像纹理信息约束及其初始化标注构建能量方程;最后,利用图割方法优化得到该能量方程的解,即为非关键帧语义标注。标注的准确率和视觉效果的实验结果表明,与运动估计法和基于模型的学习法相比较,所提基于纹理信息的室内场景语义标注学习法具有较好的效果。该方法可以为服务机器人、智能家居、应急响应等低时延决策系统提供参考。     

Learning to detect natural image boundaries using local brightness, color, and texture cues

The goal of this work is to accurately detect and localize boundaries in natural scenes using local image measurements. We formulate features that respond to characteristic changes in brightness, color, and texture associated with natural boundaries. In order to combine the information from these features in an optimal way, we train a classifier using human labeled images as ground truth. The output of this classifier provides the posterior probability of a boundary at each image location and orientation. We present precision-recall curves showing that the resulting detector significantly outperforms existing approaches. Our two main results are 1) that cue combination can be performed adequately with a simple linear model and 2) that a proper, explicit treatment of texture is required to detect boundaries in natural images.     

Median-type filters with model-based preselection masks

To remove impulsive noise, several variants of median-type filters are presented. They use some prior information provided by the probabilistic composition model of the class of the image being processed. This information is doubly applied; first, in elaborating a noise-estimation mask for the preselection of the pixels which must be filtered; and second, in the filtering process itself. The aim is to avoid the side-effect damage that median-type filters produce in processing the non-contaminated pixels. Experimental results are better than in the conventional median filter.     

Intensity contrast masks for gender classification

This work develops effective masks for AdaBoost-based gender classification. A set of intensity contrast masks around small regions of two neighbor pixels is proposed. The design exploits the sharp intensity contrast that is common to the male face and the smooth intensity contrast that is typically exhibited by the female face. The number of weak classifiers is only approximately 0.5 % of the set of weak classifiers based on pixel comparisons that were proposed by S. Baluja and H. Rowley (Int J Comput Vis 71(1):111–119, 2007). Experimental results demonstrate the feasibility of the proposed approach, and confirm the basic ideas on which this study is based. The first is that intensity contrast in face images is essential to gender classification, and the number of intensity contrast features are more required than are those intensity contrast features designed based on pixel comparisons. Second, structural information in face images may not enable features to be distinguished. Third, although intensity contrast is important to gender classification, other factors should be considered in designing a complete set of gender-related features.     

X-Ray masks for very deep X-Ray lithography

 The high aspect ratio, deep x-ray lithography and electrodeposition process [Becker et al. (1986)] can be expensive unless throughput is high enough. The use of a very high energy synchrotron has allowed the cost of exposure to be significantly reduced through simultaneous exposure of stacked photoresist [Guckel et al (1994)]. Synchrotron radiation at high photon energies has resulted the use of a large area x-ray mask. Both stacked exposures and a large area x-ray masks have significantly increased the throughput of the deep x-ray lithography and electrodeposition process. Received: 25 August 1997/Accepted: 3 September     

Precision of micromilled x-ray masks and exposures

 Micromechanical milling has been shown to be a rapid and direct method for the fabrication of structures with the geometry and size suitable for use as x-ray mask absorbers. While the micromilling process can not duplicate the size and resolution of absorber patterns created by high energy electron beam or optical lithography methods, micromilling can repeatedly create absorber line widths down to 10 micrometers, or less, with a one-sigma tolerance of 0.5 micrometers. A method for easily characterizing milling tool run out has been adapted so tool change out can be more routine. The milling process leaves some absorber burrs and the absorber is apparently tapered at the machined wall which introduce process biases, both of which add to exposure degradation. Nevertheless, based on work to date, it appears both of these effects can be reduced to acceptable limits. Received: 25 August 1997/Accepted: 3 September 1997     

基于分数阶微分的图像增强模板

根据二维数字图像具有自相关性,为了充分利用邻近像素点的信息,推导出基于分数阶Riemann-Liouville定义的模板系数,构造了八个方向的分数阶图像增强模板;同时引进信息论中熵的概念对图像增强后的纹理保留效果进行定量分析。实验表明,提出的分数阶微分图像增强模板与传统方法相比具有更好的增强效果,并有效保留了图像的纹理细节信息,对纹理具有特殊需求的应用具有一定意义。