Automatic Feature Detection Research Articles

COMPARISON AND ASSESSMENT OF 3D REGISTRATION AND GEOREFERENCING APPROACHES OF POINT CLOUDS IN THE CASE OF EXTERIOR AND INTERIOR HERITAGE BUILDING RECORDING

Abstract. In the field of 3D heritage documentation, point cloud registration is a relatively common issue. With rising needs for Historic Building Information Models (HBIMs), this issue has become more important as it determines the quality of the data to be used for HBIM modelling. Furthermore, in the context of historical buildings, it is often interesting to document both the exterior façades as well as the interior. This paper will discuss two approaches of the registration and georeferencing of building exterior and interior point clouds coming from different sensors, namely the independent georeferencing method and the free-network registration and georeferencing. Building openings (mainly windows) were used to establish common points between the systems. These two methods will be compared in terms of geometrical quality, while technical problems in performing them will also be discussed. Furthermore, an attempt to automate some parts of the workflow using automatic 3D keypoints and features detection and matching will also be described in the paper. Results show that while both approaches give similar results, the independent approach requires less work to perform. However, the free-network method has the advantage of being able to compensate for any systematic georeferencing error on either system. As regards to the automation attempt, the use of 3D keypoints and features may reduce processing time; however correct tie point correspondence filtering remains difficult in the presence of heavy point cloud noise.

Detecting Neolithic Burial Mounds from LiDAR-Derived Elevation Data Using a Multi-Scale Approach and Machine Learning Techniques

Airborne LiDAR technology is widely used in archaeology and over the past decade has emerged as an accurate tool to describe anthropomorphic landforms. Archaeological features are traditionally emphasised on a LiDAR-derived Digital Terrain Model (DTM) using multiple Visualisation Techniques (VTs), and occasionally aided by automated feature detection or classification techniques. Such an approach offers limited results when applied to heterogeneous structures (different sizes, morphologies), which is often the case for archaeological remains that have been altered throughout the ages. This study proposes to overcome these limitations by developing a multi-scale analysis of topographic position combined with supervised machine learning algorithms (Random Forest). Rather than highlighting individual topographic anomalies, the multi-scalar approach allows archaeological features to be examined not only as individual objects, but within their broader spatial context. This innovative and straightforward method provides two levels of results: a composite image of topographic surface structure and a probability map of the presence of archaeological structures. The method was developed to detect and characterise megalithic funeral structures in the region of Carnac, the Bay of Quiberon, and the Gulf of Morbihan (France), which is currently considered for inclusion on the UNESCO World Heritage List. As a result, known archaeological sites have successfully been geo-referenced with a greater accuracy than before (even when located under dense vegetation) and a ground-check confirmed the identification of a previously unknown Neolithic burial mound in the commune of Carnac.

Deep Learning Nuclei Detection in Digitized Histology Images by Superpixels

Background: Advances in image analysis and computational techniques have facilitated automatic detection of critical features in histopathology images. Detection of nuclei is critical for squamous epithelium cervical intraepithelial neoplasia (CIN) classification into normal, CIN1, CIN2, and CIN3 grades. Methods: In this study, a deep learning (DL)-based nuclei segmentation approach is investigated based on gathering localized information through the generation of superpixels using a simple linear iterative clustering algorithm and training with a convolutional neural network. Results: The proposed approach was evaluated on a dataset of 133 digitized histology images and achieved an overall nuclei detection (object-based) accuracy of 95.97%, with demonstrated improvement over imaging-based and clustering-based benchmark techniques. Conclusions: The proposed DL-based nuclei segmentation Method with superpixel analysis has shown improved segmentation results in comparison to state-of-the-art methods.

Object-Based Image Analysis: A Review of Developments and Future Directions of Automated Feature Detection in Landscape Archaeology

Object based image analysis (OBIA) is a method of assessing remote sensing data that uses morphometric and spectral parameters simultaneously to identify features in remote sensing imagery. Over the past 15-20 years, OBIA methods have been introduced to detect archaeological features. Improvements in accuracy have been attained by using a greater number of morphometric variables and multiple scales of analysis. This article highlights the developments that have occurred in the application of OBIA within archaeology and traces some possible paths for the future uses of these methods. I argue that OBIA is both a useful and necessary tool for archaeology, and discuss future research paths using this method. Some of the suggestions put forth here include: pushing for multifaceted research designs utilizing OBIA and manual interpretation, using OBIA methods for directly studying landscape settlement patterns, and increasing data sharing of methods between researchers.

The contribution of pattern recognition techniques in geomorphology and geology: the case study of Tinos Island (Cyclades, Aegean, Greece)

ABSTRACTScope of the present work is to apply modern methods of pattern recognition concerning the automatic detection of geomorphological features (curvilinear lineaments and topographic highs), with emphasis on geological faults and to compare the results of the automatic detection with aerial photographs and further geological data collected in Tinos Island. The contribution of this work in the geosciences is multidisciplinary because (a) it proposes to geoscientists a new tool of geomorphological analysis prior to the fieldwork and especially at the stage where published work and already available data are collected, evaluated and reprocessed, (b) it increases the detection accuracy of the geological features both prior and after the fieldwork, (c) it decreases the time of fieldwork and consequently the cost of the entire geological research and (d) it can be applied in digital elevation data. The automatically detected linear-curvilinear lineaments and topographic highs agree in location, shape and orientation with the ground truth data (geological maps, aerial photographs and field measurements). Furthermore, the shape and the orientation of the most prominent topographic high (Tsiknias) seems to be related to the tectonic regime in the wide area of study corresponding to the ENE-WSW, WSW-ESE, N-S and E-W trending normal faults.

Automatic detection of solar features in HSOS full-disk solar images using guided filter

A procedure is introduced for the automatic detection of solar features using full-disk solar images from Huairou Solar Observing Station (HSOS), National Astronomical Observatories of China. In image preprocessing, median filter is applied to remove the noises. Guided filter is adopted to enhance the edges of solar features and restrain the solar limb darkening, which is first introduced into the astronomical target detection. Then specific features are detected by Otsu algorithm and further threshold processing technique. Compared with other automatic detection procedures, our procedure has some advantages such as real time and reliability as well as no need of local threshold. Also, it reduces the amount of computation largely, which is benefited from the efficient guided filter algorithm. The procedure has been tested on one month sequences (December 2013) of HSOS full-disk solar images and the result shows that the number of features detected by our procedure is well consistent with the manual one.

Testing Google Earth Engine for the automatic identification and vectorization of archaeological features: A case study from Faynan, Jordan

Google Earth Engine (GEE) is an in-development, cloud-based platform providing access to petabytes of satellite imagery data for planetary-scale analysis (Google Earth Engine Team 2015). Combining this massive database with the parallel computing power of Google's infrastructure facilitates quick and easy analysis of satellite imagery on any scale, opening new avenues for research in a number of fields. This paper evaluates the potential role GEE can play in the future of archaeological research. To do so, GEE was employed/tested in two case studies. First, GEE was used to automatically identify specific archaeological features across the landscape of the archaeologically-rich Faynan region of Southern Jordan. Second, GEE-based edge-detection and automatic vectorization for mapping archaeological sites was tested at the Iron Age (ca. 1200–900BCE) site of Khirbat al-Jariya in Faynan. Based on the test results, the authors concluded that GEE has significant potential for assisting archaeologists with automated feature detection and vectorization, tasks that are often onerous and expensive.

Robust Detection and Visualization of Jet-Stream Core Lines in Atmospheric Flow.

Jet-streams, their core lines and their role in atmospheric dynamics have been subject to considerable meteorological research since the first half of the twentieth century. Yet, until today no consistent automated feature detection approach has been proposed to identify jet-stream core lines from 3D wind fields. Such 3D core lines can facilitate meteorological analyses previously not possible. Although jet-stream cores can be manually analyzed by meteorologists in 2D as height ridges in the wind speed field, to the best of our knowledge no automated ridge detection approach has been applied to jet-stream core detection. In this work, we -a team of visualization scientists and meteorologists-propose a method that exploits directional information in the wind field to extract core lines in a robust and numerically less involved manner than traditional 3D ridge detection. For the first time, we apply the extracted 3D core lines to meteorological analysis, considering real-world case studies and demonstrating our method's benefits for weather forecasting and meteorological research.

Recognition of feature curves on 3D shapes using an algebraic approach to Hough transforms

Feature curves are largely adopted to highlight shape features, such as sharp lines, or to divide surfaces into meaningful segments, like convex or concave regions. Extracting these curves is not sufficient to convey prominent and meaningful information about a shape. We have first to separate the curves belonging to features from those caused by noise and then to select the lines, which describe non-trivial portions of a surface. The automatic detection of such features is crucial for the identification and/or annotation of relevant parts of a given shape. To do this, the Hough transform (HT) is a feature extraction technique widely used in image analysis, computer vision and digital image processing, while, for 3D shapes, the extraction of salient feature curves is still an open problem.Thanks to algebraic geometry concepts, the HT technique has been recently extended to include a vast class of algebraic curves, thus proving to be a competitive tool for yielding an explicit representation of the diverse feature lines equations. In the paper, for the first time we apply this novel extension of the HT technique to the realm of 3D shapes in order to identify and localize semantic features like patterns, decorations or anatomical details on 3D objects (both complete and fragments), even in the case of features partially damaged or incomplete. The method recognizes various features, possibly compound, and it selects the most suitable feature profiles among families of algebraic curves.

An approach for behavior analysis using correlation spectral embedding method

Automatic identification of various facial movements and expressions with high recognition value is important for human computer interaction as the facial behavior of a human can be treated as an important factor for information representation as well as communication. A high deviation of human appearance and existence of noisy contextual background makes the human pose analysis is hard to achieve. A number of basic factors such as cluttered background, occlusion, and camera movement and illumination variations degrade the image quality resulting in poor performance for identifying different facial expressions. Moreover, the identification of the automatic feature detection in facial behavior requires high degree of correlation between the training and test images. Our proposed work tries to address the mentioned problems and resolve to some extent. In this methodology, a Decision-based Spectral Embedding approach combining appearance and geometry based features for head pose estimation and facial expression recognition by minimizing the objective function which leads to selection of optimal set of fiducial points. The method preserves the local information from different facial views for mapping neighboring input to its corresponding output, resulting in low dimensional representation for encoding the relationships of the data. The proposed methodology is validated with benchmark datasets for analyzing the performance of recognition of facial behavior.

Curvature-Based Feature Detection for Head Modeling

In the field of 3D head modeling and animation, feature points are often needed to mark important regions of the face and can be used to animate or deform the input model. However, an automatic detection of features remains a challenging task. This paper presents a novel approach to feature detection based on curvature and its derived descriptors, such as shape index, curvedness and Willmore energy. Four important feature regions are detected using the proposed approach - eyes, nose, mouth, ears. For each region, feature points are detected. Results show that the feature points are detected with sufficient accuracy for further use.

Exudate and Blood Vessel Feature Extraction in Diabetic Retinopathy Patients using Morphology Operation

Diabetic Retinopathy is one of the retina complications caused by diabetic disease with observable symptoms such as emergence of exudate and new blood vessels. The tool used to screen it is a fundus camera. However, analyzing the fundus image should be done by doctor who is an expert and will require a lot of time. Therefore, automatic feature detection can assist doctor in processing the retinal image in analyzing diabetic retinopathy disease. The proposed method has been tested on the morphological operations of the fundus image from Cicendo Eye Hospital, Bandung. The calculation results on feature extraction exudate area has a range of 0 pixels for normal retinal image, 17-21213 pixel for retinal image NPDR, and 125-12299 retinal image pixel for PDR. The calculation results on the extraction area of blood vessels has a range of 13319-46681 pixel to the normal retina, the retinal image 7435-49938 pixel for NPDR, and 13.81-53.802 retinal image pixel for PDR.

Hexagonal Descriptor Particle Swarm Optimization with Knowledge-Crowding for Face Recognition

Automatic identification of various facial expressions with high recognition value is important for human computer interaction as the facial behavior of a human can be treated as an important factor for information representation as well as communication. A number of basic factors such as cluttered background, occlusion, and camera movement and illumination variations degrade the image quality resulting in poor performance for identifying different facial expressions. Moreover the identification of the automatic feature detection in facial behavior requires high degree of correlation between the training and test images. Face recognition is done by minimizing the objective function which leads to selection of optimal set of fiducial points. The method preserves the local information from different facial views for mapping neighboring input to its corresponding output, resulting in low dimensional representation for encoding the relationships of the data. The proposed method Hexagonal Descriptor Particle Swarm Optimization with Knowledge-Crowding (HDPSO-KC) overcomes from local optima and improves global search process and collaborative work. The method also covers the problem of eliminating the particles in denser regions in Pareto front distribution. The proposed methodology is validated with benchmark datasets for analyzing the performance over other methods.

Classification of malignant melanoma and benign skin lesions: implementation of automatic ABCD rule

The ABCD (asymmetry, border irregularity, colour and dermoscopic structure) rule of dermoscopy is a scoring method used by dermatologists to quantify dermoscopy findings and effectively separate melanoma from benign lesions. Automatic detection of the ABCD features and separation of benign lesions from melanoma could enable earlier detection of melanoma. In this study, automatic ABCD scoring of dermoscopy lesions is implemented. Pre-processing enables automatic detection of hair using Gabor filters and lesion boundaries using geodesic active contours. Algorithms are implemented to extract the characteristics of ABCD attributes. Methods used here combine existing methods with novel methods to detect colour asymmetry and dermoscopic structures. To classify lesions as melanoma or benign nevus, the total dermoscopy score is calculated. The experimental results, using 200 dermoscopic images, where 80 are malignant melanomas and 120 benign lesions, show that the algorithm achieves 91.25% sensitivity of 91.25 and 95.83% specificity. This is comparable to the 92.8% sensitivity and 90.3% specificity reported for human implementation of the ABCD rule. The experimental results show that the extracted features can be used to build a promising classifier for melanoma detection.

Change detection of linear features in temporally spaced remotely sensed images using edge-based grid analysis

Automatic change detection of land cover features using high-resolution satellite images, is a challenging problem in the field of intelligent remote sensing data interpretation, and is becoming more and more effective for its applications viz. urban planning and monitoring, disaster assessment etc. In the present study, a change in detection approach based on the image morphology that analyses change in the local image grids is proposed. In this approach, edges from both the images are extracted and grid wise comparison is made by probabilistic thresholding and power spectral density analysis for identifying change area. One of the advantages of the proposed methodology is that the temporal images used in the change analysis need not be radiometrically corrected as analysis is based on edge extractions. The grid-based analysis further reduces the error, which might have been introduced by image mis-registration. The proposed methodology is validated by finding the temporal changes in the linear land cover features in parts of Kolkata city, India using three different image data-sets from LISS IV, Cartosat-1 and Google earth having varied spatial resolutions of 5.8 m, 2.5 m and about 1 m, respectively. The overall accuracy in identifying changes is found to be 64.82, 73.86 and 80.93% for LISS IV, Cartosat-1 and Google earth data-set, respectively.

Feature detection and matching on atmospheric nuclear detonation video

Automated feature matching of nuclear detonations (NUDETs) enables three-dimensional point cloud reconstruction, and establishment a volume-based model to reduce uncertainty in estimating the yield of NUDETs solely from video. Establishing a volume-based model requires feature correspondences between wide viewpoints of 58°–110° that are larger than scale-invariant feature transform-based techniques can reliably match. The presented technique detects relative bright features in the NUDET known as ‘hotspots,’ and matches them across wide viewpoints using a spherical based object model. Results show that hotspots can be detected with a 71.95% hit rate and 86.03% precision. Hotspots are matched to films from different viewpoints with 76.6% correctness and a standard deviation of 16.4%. Hotspot descriptors are also matched in time sequence with 99.6% correctness and a standard deviation of 1.07%. The results demonstrate that a spherical model can serve as a viable descriptor model for matching across wide viewpoints when the object is known to be spherical. It also demonstrates an automated feature detection and matching combination that enables features to be matched from unsynchronised video across wide viewpoints of 58°–110° on spherical objects where state-of-the-art techniques are insufficient.

Automatic Detection of Geometric Features in CAD models by Characteristics

ABSTRACTIn this paper, we detailed a novel 3D detection method for geometric features in CAD models used in CAD-to-CAE model simplification process. Model preparation is now the most time consuming part in the computer-aided engineering process chain, it is commonly labor intensive. Therefore, the overall objectives are: 1) introduce automation to speed-up the model preparation process; 2) remove redundant features to reduce model mesh size for fast mesh generation and analysis without compromising on solution accuracy. Automatic feature detection that is generic and easily extendable is a key element needed to achieve these objectives. Inside the article, we proposed detection by characteristics to complement existing geometric modeling kernels for CAD systems with defeaturing functions to detect features previously not detectable.

Magnetic Resonance Imaging–Guided versus Surrogate-Based Motion Tracking in Liver Radiation Therapy: A Prospective Comparative Study

This study applied automatic feature detection on cine-magnetic resonance imaging (MRI) liver images in order to provide a prospective comparison between MRI-guided and surrogate-based tracking methods for motion-compensated liver radiation therapy. In a population of 30 subjects (5 volunteers plus 25 patients), 2 oblique sagittal slices were acquired across the liver at high temporal resolution. An algorithm based on scale invariant feature transform (SIFT) was used to extract and track multiple features throughout the image sequence. The position of abdominal markers was also measured directly from the image series, and the internal motion of each feature was quantified through multiparametric analysis. Surrogate-based tumor tracking with a state-of-the-art external/internal correlation model was simulated. The geometrical tracking error was measured, and its correlation with external motion parameters was also investigated. Finally, the potential gain in tracking accuracy relying on MRI guidance was quantified as a function of the maximum allowed tracking error. An average of 45 features was extracted for each subject across the whole liver. The multi-parametric motion analysis reported relevant inter- and intrasubject variability, highlighting the value of patient-specific and spatially-distributed measurements. Surrogate-based tracking errors (relative to the motion amplitude) were were in the range 7% to 23% (1.02-3.57 mm) and were significantly influenced by external motion parameters. The gain of MRI guidance compared to surrogate-based motion tracking was larger than 30% in 50% of the subjects when considering a 1.5-mm tracking error tolerance. Automatic feature detection applied to cine-MRI allows detailed liver motion description to be obtained. Such information was used to quantify the performance of surrogate-based tracking methods and to provide a prospective comparison with respect to MRI-guided radiation therapy, which could support the definition of patient-specific optimal treatment strategies.

Use of Gestalt Theory and Random Sets for Automatic Detection of Linear Geological Features

This paper presents the calibration and application of a Gestalt-based line segment method for automatic geological lineament detection from remote sensing images. This method involves estimation of the scale factor, the angle tolerance and a threshold on the false alarm rate. It identifies major lineaments as objects characterized by two edges on the image, which appear as transitions from dark to bright and vice versa. These objects were modelled as random sets with parameters drawn from their distributions. Following the geometry of detected segments, a novel validation method assesses the accuracy with respect to a linear vector reference. The methodology was applied to a study area in Kenya where lineaments are prominent in the landscape and are well identifiable from an ASTER image. Error rates were based on distance and local orientation, and the study showed that the existence and size of the objects were sensitive to parameter variation. False detection rate and missing detection rate were both equal to 0.50, which is better than values equal to 0.65 and 0.63, observed using the Canny edge detection. Modelling the uncertainty of geological lineaments with random sets further showed that no core set is formed, indicating that there is an inherent uncertainty in their existence and position, and that the variance is relatively high. Comparing the test area with four areas in the same region showed similar results. Despite some shortcomings in identifying full lineaments from partially observed lineaments, it is concluded that the procedure in this paper is well able to automatically extract lineaments from a remote sensing image and validate their existence.

Fully Automatic Detection of Salient Features in 3-D Transesophageal Images

Most automated segmentation approaches to the mitral valve and left ventricle in 3-D echocardiography require a manual initialization. In this article, we propose a fully automatic scheme to initialize a multicavity segmentation approach in 3-D transesophageal echocardiography by detecting the left ventricle long axis, the mitral valve and the aortic valve location. Our approach uses a probabilistic and structural tissue classification to find structures such as the mitral and aortic valves; the Hough transform for circles to find the center of the left ventricle; and multidimensional dynamic programming to find the best position for the left ventricle long axis. For accuracy and agreement assessment, the proposed method was evaluated in 19 patients with respect to manual landmarks and as initialization of a multicavity segmentation approach for the left ventricle, the right ventricle, the left atrium, the right atrium and the aorta. The segmentation results revealed no statistically significant differences between manual and automated initialization in a paired t-test (p > 0.05). Additionally, small biases between manual and automated initialization were detected in the Bland–Altman analysis (bias, variance) for the left ventricle (−0.04, 0.10); right ventricle (−0.07, 0.18); left atrium (−0.01, 0.03); right atrium (−0.04, 0.13); and aorta (−0.05, 0.14). These results indicate that the proposed approach provides robust and accurate detection to initialize a multicavity segmentation approach without any user interaction.