Orientation-based Method Research Articles

Efficient Descriptor of Histogram of Ridges Orientation Delineate for Fingernail

Fingernails structure are rich in orientation, ridges and edge features. Inspired by Edge Histogram Descriptor (EHD), this paper presents an efficient orientation-based local descriptor, named histogram of ridges orientation delineate (HROD). HROD is based on the fact that human vision is sensitive to edge features for image perception. For a given image, HROD algorithm first execute and perform a pre-process i.e., re-sizing, filtering, enhancement, segmentation, edge detection and feature extraction. Then, finds oriented edge maps according to predefined orientations using a well-known edge operator mask (2×2 sub block) and obtains a ridges orientation delineate map by choosing an orientation with the maximum edge magnitude for each pixel. In the experiment on this research, five oriented edge maps were used to generate and detect the maximum edge orientation construction of each block, namely vertical, horizontal, diagonal 45°, diagonal 135° and isotropic (non-orientation specific) orientation. Experimental results on fingernail images show that the performance of HROD comparable with the state-of-the-art orientation-based methods (e.g., Gabor filter, histogram of oriented gradients, and local directional code). Furthermore, the proposed HROD algorithm has advantages of low feature dimensionality and fast implementation for a real-time fingernails orientation recognition system.

Comprehensive identification of statistical homogeneity of fractured rock masses for a candidate HLW repository site, China

Determining the statistically homogeneous regions is important for engineering geological modeling and promoting systematic research on regional rock mass characteristics and rock engineering designs. This study proposes a comprehensive framework considering multiple discontinuity geometry parameters (e.g., orientation, trace length, density, and cluster characteristics.) to distinguish statistical homogeneity of fractured rock masses. First, based on the equal-area plot method (Schmidt net), the quantitative judgment ability of the Miller method for determining the similarity between two discontinuity orientation populations was improved by optimizing block network division scheme and blank area merging principle. Then, considering the distribution characteristics of discontinuity in rock mass space and the variation features of discontinuity at different geographical locations, a three-dimensional fracture network model, established using the discontinuity geometry parameters obtained from field investigation, was combined with the fractal dimension method (called the network fractal dimension method; NFD) to further evaluate the statistical homogeneity of fractured rock mass. The proposed approaches were subsequently applied to the identification of the statistical homogeneity of several typical granite fracture outcrops in the Beishan area, which is a candidate site for high-level radioactive waste (HLW) repository of China. The case study reveals that the optimized Miller method is more flexible and applicable than the conventional orientation-based methods, especially in dealing with the complex situation of the coexistence of gentle dip angle and steep dip angle as well as clustering characteristics of discontinuity. The NFD method was used to unify the influence of the main discontinuity characteristic parameters for a logical and unbiased evaluation of the statistical homogeneity. The comprehensive framework developed herein integrates the advantages of the two methods and reduces the interference of subjective evaluation. Therefore, it can be employed to efficiently determine the statistical homogeneity of rock masses at different locations at a large scale without concern over the deviation caused by different directions and size of the compared areas.

A revised orientation-based correction method for SfM-MVS point clouds of outcrops using ground control planes with marks

Structure-from-Motion/Multiview Stereo (SfM-MVS) point cloud has the potential to display objects and to generate corresponding digital elevation model (DEM) and orthomosaic, from which geometric parameters such as the orientations, shapes, and sizes can be derived, thus making it a promising tool for structural geology. High-precision coordinates of ground control points (GCP) are commonly used for SfM-MVS point cloud correction, but the coordinate measurement depends on expensive and heavy equipment, and more importantly, there is uncertainty of the Global Position System (GPS) signal in some environments. A potential alternative correction method is to use orientations from outcrops, although this method still requires further optimization. This study aims to improve this orientation-based method using ground control planes (GCPL) and test its effectiveness through geometric analysis and simulation experiments, and the improved schemes are as follows: using a single pair of nonparallel planes, using a single horizontal plane with linear orientation marks, and using a single inclined plane with a strike mark. The GCPL method requires a few orientations, and its accuracy is sufficient for fieldwork. It greatly simplifies the field investigation and improves the applicability of SfM–MVS in environments with unavailable GPS signals or with rigid body rotations for GPS imprecision.

A Progressive Framework for Delineating Homogeneous Domains in Complicated Fractured Rock Masses: A Case Study from the Xulong Dam Site, China

As components of natural geological bodies, rock masses generally experience long-term endogenic and exogenic geological processes. Since the composition and structure of rock masses generally vary from one domain to another, it is necessary to divide rock masses into several approximately homogeneous domains before developing rock engineering designs. A progressive framework that integrates geological, geotechnical and structural aspects was proposed to demarcate homogeneous domains. The proposed framework was applied to identify the homogeneous domains along three exploration tunnels at the Xulong dam site. The demarcation of geological domains was mainly done by means of thin-section examination of rock specimens and outcrop investigation. Field observations and P wave velocity tests were used to distinguish the changes in the geotechnical properties of rock masses caused by rock unloading and weathering. The crack tensor, which can synthetically consider the orientation, dimension and volume density of discontinuities, was introduced to determine the structural similarity of rock masses based on in situ measurable quantities. The results show that the crack tensor can effectively distinguish structural differences and exhibits better performance than conventional orientation-based methods. A major advantage of the proposed progressive framework is that it provides a quantitative, logical and reasonable framework to delineate homogeneous domains based on the comprehensive utilization of available field data in a logical sequence.

Investigating local orientation methods to segment microstructure with 3D solid texture

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.

Palmprint Recognition Using Neighboring Direction Indicator

Orientation features are successfully used in coding-based palmprint recognition methods. In this paper, we propose a discriminative neighboring direction indicator to represent the orientation feature of the palmprint. The neighboring direction indicator feature not only represents the most dominant orientation feature of the palmprint, but also better describes the orientation feature of those points which have double dominant orientations. In addition, the neighboring direction indicator shows good robustness to noise and rotation. Using the neighboring direction indicator, we propose a novel palmprint recognition method. Extensive experiments conducted on three types of palmprint databases demonstrate that the proposed method gives better performance than the existing state-of-the-art orientation-based methods. By using the proposed method, the equal error rate is improved by about 10% for palmprint verification, and the average error rate is reduced by 2.7–14% for palmprint identification with a single training sample.

Efficient descriptor of histogram of salient edge orientation map for finger vein recognition.

Finger vein images are rich in orientation and edge features. Inspired by the edge histogram descriptor proposed in MPEG-7, this paper presents an efficient orientation-based local descriptor, named histogram of salient edge orientation map (HSEOM). HSEOM is based on the fact that human vision is sensitive to edge features for image perception. For a given image, HSEOM first finds oriented edge maps according to predefined orientations using a well-known edge operator and obtains a salient edge orientation map by choosing an orientation with the maximum edge magnitude for each pixel. Then, subhistograms of the salient edge orientation map are generated from the nonoverlapping submaps and concatenated to build the final HSEOM. In the experiment of this paper, eight oriented edge maps were used to generate a salient edge orientation map for HSEOM construction. Experimental results on our available finger vein image database, MMCBNU_6000, show that the performance of HSEOM outperforms that of state-of-the-art orientation-based methods (e.g., Gabor filter, histogram of oriented gradients, and local directional code). Furthermore, the proposed HSEOM has advantages of low feature dimensionality and fast implementation for a real-time finger vein recognition system.

An Effective Orientation-based Method and Parameter Space Discretization for Defined Object Segmentation

While non-predefined object segmentation (NDOS) distinguishes an arbitrary self-assumed object from its background, predefined object segmentation (DOS) pre-specifies the target object. In this paper, a new and novel method to segment predefined objects is presented, by globally optimizing an orientation-based objective function that measures the fitness of the object boundary, in a discretized parameter space. A specific object is explicitly described by normalized discrete sets of boundary points and corresponding normal vectors with respect to its plane shape. The orientation factor provides robust distinctness for target objects. By considering the order of transformation elements, and their dependency on the derived over-segmentation outcome, the domain of translations and scales is efficiently discretized. A branch and bound algorithm is used to determine the transformation parameters of a shape model corresponding to a target object in an image. The results tested on the PASCAL dataset show a considerable achievement in solving complex backgrounds and unclear boundary images.

Logic-based topological representation of vague moving regions: computational models for well-behaved GIS solutions

This study proposes a solution for the topological representation and interpretation of moving regions under vagueness in a geographic information system (GIS). The problem is multifaceted, and different aspects will be discussed. We investigate the impact of several types of fuzzy logic on results and address which is more convenient for the region connection calculus under fuzziness (FRCC). By recognising that the main and basic relation to evaluate topological relations is the connection relation, we adopt a qualitative strategy based on fuzzy inference and resemblance relations. We also define topological differences and introduce fuzzy transition relations (FTRs) to distinguish how a transition from one topological situation to another occurs for a pair of objects during a time period. As pure topological relations are not sufficient for a complete qualitative perception, an orientation-based method (OBM) will be proposed. Using this method, we can interpret, for example, a region that partially overlaps another region mostly on its northern part. Most importantly, some step-by-step algorithms and constructs are proposed and evaluated such that the model can apply to regions stored in vector data models.

Bottom-up estimation of joint moments during manual lifting using orientation sensors instead of position sensors

L5/S1, hip and knee moments during manual lifting tasks are, in a laboratory environment, frequently established by bottom-up inverse dynamics, using force plates to measure ground reaction forces (GRFs) and an optoelectronic system to measure segment positions and orientations. For field measurements, alternative measurement systems are being developed. One alternative is the use of small body-mounted inertial/magnetic sensors (IMSs) and instrumented force shoes to measure segment orientation and GRFs, respectively. However, because IMSs measure segment orientations only, the positions of segments relative to each other and relative to the GRFs have to be determined by linking them, assuming fixed segment lengths and zero joint translation. This will affect the estimated joint positions and joint moments. This study investigated the effect of using segment orientations only (orientation-based method) instead of using orientations and positions (reference method) on three-dimensional joint moments. To compare analysis methods (and not measurement methods), GRFs were measured with a force plate and segment positions and/or orientations were measured using optoelectronic marker clusters for both analysis methods. Eleven male subjects lifted a box from floor level using three lifting techniques: a stoop, a semi-squat and a squat technique. The difference between the two analysis methods remained small for the knee moments: <4%. For the hip and L5/S1 moments, the differences were more substantial: up to 8% for the stoop and semi-squat techniques and up to 14% for the squat technique. In conclusion, joint moments during lifting can be estimated with good accuracy at the knee joint and with reasonable accuracy at the hip and L5/S1 joints using segment orientation and GRF data only.