Contour Finding Algorithm Research Articles

A comparison of past and present computational methods for shape analysis of double-shell x-ray radiographs.

Implosion symmetry is a key requirement in achieving a robust burning plasma in inertial confinement fusion experiments. In double-shell capsule implosions, we are interested in the shape of the inner shell as it pushes on the fuel. Shape analysis is a popular technique for studying said symmetry during implosion. Combinations of filtering and contour-finding algorithms are studied for their promise in reliably recovering Legendre shape coefficients from synthetic radiographs of double-shell capsules with applied levels of noise. A radial lineout max(slope) method when used on an image pre-filtered with non-local means and a variant of the marching squares algorithm are able to recover p0, p2, and p4 maxslope Legendre shape coefficients with mean pixel discrepancy errors of 2.81 and 3.06, respectively, for the noisy synthetic radiographs we consider. This improves upon prior radial lineout methods paired with Gaussian filtering, which we show to be unreliable and whose performance is dependent on input parameters that are difficult to estimate.

A dual U-Net algorithm for automating feature extraction from satellite imagery

Quick identification of building and roads is critical for execution of tactical US military operations in an urban environment. To this end, a gridded, referenced, satellite images of an objective, often referred to as a gridded reference graphic or GRG, has become a standard product developed during intelligence preparation of the environment. At present, operational units identify key infrastructure by hand through the work of individual intelligence officers. Recent advances in Convolutional Neural Networks, however, allows for this process to be streamlined through the use of object detection algorithms. In this paper, we describe an object detection algorithm designed to quickly identify and label both buildings and road intersections present in an image. Our work leverages both the U-Net architecture as well the SpaceNet data corpus to produce an algorithm that accurately identifies a large breadth of buildings and different types of roads. In addition to predicting buildings and roads, our model numerically labels each building by means of a contour finding algorithm. Most importantly, the dual U-Net model is capable of predicting buildings and roads on a diverse set of test images and using these predictions to produce clean GRGs.

Monocular positioning of sweet peppers: An instance segmentation approach for harvest robots

Accurate positioning of fruit is a key issue that has attracted much attention in the field of harvest robots. The complex environment and close proximity make the perception of dense crops in greenhouses a challenging problem. Different from various solutions proposed involving special equipment or other auxiliary information, we propose a novel positioning approach based on instance segmentation using a monocular RGB camera. To achieve high position accuracy, we first design a deep convolutional neural network (CNN) in a multi-task framework to export a binary segmentation map and an embedded feature map. To solve the problem of performance degradation in the intersection-over-union (IoU) for the binary segmentation task caused by multi-task optimisation, the encoder part of our network is redesigned on the basis of a Visual Geometry Group network with 16 convolutional layers (VGG-16). Then, mean-shift clustering is used to achieve instance segmentation. Finally, a contour-finding algorithm is presented for outlining fruit without the help of any contextual information. Based on these contours, the five fruits with the largest contour areas are selected as the targets for positioning. We verify our method on a public sweet pepper dataset and achieve competitive results. Divided by the radius of the fruit, the average position error for the first target in the harvesting order is 0.18, which shows that our method outperforms the semantic segmentation method. For the first five targets in the harvesting order, this index is less than 0.3 on average, similar to that of the semantic segmentation method with only one target output.

A Semi-automated Computational Approach for Infrared Dark Cloud Localization: A Catalog of Infrared Dark Clouds

The field of computer vision has greatly matured in the past decade, and many of the methods and techniques can be useful for astronomical applications. One example is in searching large imaging surveys for objects of interest, especially when it is difficult to specify the characteristics of the objects being searched for. We have developed a method using contour finding and convolution neural networks (CNNs) to search for Infrared Dark Clouds (IRDCs) in the Spitzer Galactic plane survey data. IRDCs can vary in size, shape, orientation, and optical depth, and are often located near regions with complex emission from molecular clouds and star formation, which can make the IRDCs difficult to reliably identify. False positives can occur in regions where emission is absent, rather than from a foreground IRDC. The contour finding algorithm we implemented found most closed figures in the mosaic and we developed rules to filter out some of the false positive before allowing the CNNs to analyze them. The method was applied to the Spitzer data in the Galactic plane surveys, and we have constructed a catalog of IRDCs which includes additional parts of the Galactic plane that were not included in earlier surveys.

Particle Swarm Contour Search Algorithm.

In this article, we present a new algorithm called Particle Swarm Contour Search (PSCS)—a Particle Swarm Optimisation inspired algorithm to find object contours in 2D environments. Currently, most contour-finding algorithms are based on image processing and require a complete overview of the search space in which the contour is to be found. However, for real-world applications this would require a complete knowledge about the search space, which may not be always feasible or possible. The proposed algorithm removes this requirement and is only based on the local information of the particles to accurately identify a contour. Particles search for the contour of an object and then traverse alongside using their known information about positions in- and out-side of the object. Our experiments show that the proposed PSCS algorithm can deliver comparable results as the state-of-the-art.

A novel computerised quantification of thyroid vascularity in the differentiation of malignant and benign thyroid nodules

PurposeOnly five percent of thyroid nodules are malignant. It is important to find reliable and at the same time non-invasive methods to identify high-risk nodules. The aim of this study was to determine the diagnostic validity of a morphologic feature-oriented approach of ultrasound study for the identification of malignant thyroid nodules.Material and methodsSeventy-one thyroid nodules in 71 consecutive patients were evaluated with both ultrasonography (US) and US-assisted fine needle aspiration biopsy (FNAB). Thyroid grey-scale and power Doppler US were performed, and a Windows-based software was designed to process power Doppler US (PDUS) images that were recorded directly by the US device. We provided a histogram graph of coloured pixels and calculated the Malignancy Index to identify the probability of malignancy for each thyroid nodule.ResultsThirty-six nodules (50.7%) were determined to be malignant in FNAB. Area under the receiver operating curve was 0.91 (95% CI: 0.85-0.98) for PDUS-based malignancy index in differentiating malignant thyroid nodules from benign ones. The best cut-off point for malignancy index was determined to be 0.092, with a sensitivity of 86.1% and specificity of 80% in identifying malignant nodules.ConclusionsThis PDUS-driven malignancy index using a contour-finding algorithm approach could accurately and reliably differentiate malignant and benign thyroid nodules. As a pre-FNAB assessment, the malignancy index may be able to reduce the number of patients with nodular thyroid disease undergoing this invasive procedure.

Development of an Android-Based Visual Sudoku Solver Using Contour Finding and Backtracking Algorithm

This study developed an Android application that automatically solves Sudoku puzzle. An image of the Sudoku puzzle is either captured by the built-in camera or loaded directly from the media gallery. Contour finding algorithm and Hough transform would detect the game board. The givens (i.e. number that exist in a Sudoku board) are acquired by dividing the entire game board by 81 cells with the contour finding algorithm extracting the particular given. Identification is done using the Tesseract – Optical Character Recognition (OCR). Backtracking algorithm calculated the solution and displayed it back in the form of an image. After testing 30 Sudoku images, a 74.47% detection percentage and 13.09% branching factor with a Receiver Operating Characteristic - Area Under the Curve (AUC) of 0.684 was obtained when identifying the givens indicating that the system is effective.

Real-Time Three-Dimensional Echocardiography of the Left Ventricle—Pediatric Percentiles and Head-to-Head Comparison of Different Contour-Finding Algorithms: A Multicenter Study

Real-time three-dimensional echocardiography (RT3DE) is a promising method for accurate assessment of left ventricular (LV) volumes and function, however, pediatric reference values are scarce. The aim of the study was to establish pediatric percentiles in a large population and to compare the inherent influence of different evaluation software on the resulting measurements. In a multicenter prospective-design study, 497 healthy children (ages 1day to 219months) underwent RT3DE imaging of the LV (ie33, Philips, Andover, MA). Volume analysis was performed using QLab 9.0 (Philips) and TomTec 4DLV2.7 (vendor-independent; testing high (TomTec75) and low (TomTec30) contour-finding activity). Reference percentiles were computed using Cole's LMS method. In 22 subjects, cardiovascular magnetic resonance imaging (CMR) was used as the reference. A total of 370/497 (74.4%) of the subjects provided adequate data sets. LV volumes had a significant association with age, body size, and gender; therefore, sex-specific percentiles were indexed to body surface area. Intra- and interobserver variability for both workstations was good (relative bias±SD for end-diastolic volume [EDV] in %: intraobserver: QLab=-0.8±2.4; TomTec30=-0.7±7.2; TomTec75=-1.9±6.7; interobserver: QLab=2.4±7.5; TomTec30=1.2±5.1; TomTec75=1.3±4.5). Intervendor agreement between QLab and TomTec30 showed larger bias and wider limits of agreement (bias: QLab vs TomTec30: end-systolic volume [ESV]=0.8%±23.6%; EDV=-2.2%±17.0%) with notable individual differences in small children. QLab and TomTec underestimated CMR values, with the highest agreement between CMR and QLab. RT3DE allows reproducible noninvasive assessment of LV volumes and function. However, intertechnique variability is relevant. Therefore, our software-specific percentiles, based on a large pediatric population, serve as a reference for both commonly used quantification programs.

Measurement of Left and Right Atrial Volume in Patients Undergoing Ablation for Atrial Arrhythmias: Comparison of a Manual versus Semiautomatic Algorithm of Real Time 3D Echocardiography

Real time full-volume 3D echocardiography (3DE) allows rapid and noninvasive measurement of left (LA) and right atrial (RA) volume without geometric assumptions. Different algorithms from different commercial providers are available. Older software requires manual tracing of endocardial contours. Recently, software with semiautomatic endocardial contour-finding algorithms has become available, which considerably speeds up the procedure. Our aim was to compare, in the same dataset, both LA and RA volumes determined by an algorithm involving manual tracing to the corresponding volumes obtained by an algorithm with semiautomatic contour detection. Maximal atrial volumes were measured in 88 patients using a multiplane interpolation method algorithm based on manual planimetry of 8 slices. These volumes were compared with volumes determined by the QLAB 8.1 software using semiautomatic border detection. Linear regression showed excellent correlation between volumes determined by manual and by semiautomatic software for both LA and RA (r(2) = 0.90 and 0.89, respectively, P < 0.001). Bland-Altman analysis of manual versus semiautomatic volume determination showed narrow 95% limits of agreement (-15.9 to +12.0 mL for LA volume and -13.9 to +12.2 mL for RA volume) with a minimal bias of -1.9 ± 7.0 mL and -0.8 ± 6.5 mL, respectively, by the semiautomatic method. The semiautomatic border detection method shows excellent correlation for maximal LA and RA volume determination compared to the more time-consuming, multiplane interpolation method, with only slight underestimation. The results indicate that values of LA and RA volumes obtained by either algorithm can be compared, for example, during follow-up examinations.

화방 정찰 체계에서의 다수의 이동 로봇을 위한 시간 효율적인 경로 계획 알고리즘에 대한 연구

Since nuclear and chemical materials could damage people and disturb battlefield missions in a wide region, nuclear/chemical reconnaissance systems utilizing multiple mobile robots are highly desirable for rapid and safe reconnaissance. In this paper, we design a nuclear/chemical reconnaissance system including mobile robots. Also we propose time-efficient trajectory planning algorithms using grid coverage and contour finding methods for reconnaissance operation. For grid coverage, we performed in analysis on time consumption for various trajectory patterns generated by straight lines and arcs. We proposed BCF (Bounded Contour Finding) and BCFEP (Bounded Contour Finding with Ellipse Prediction) algorithms for contour finding. With these grid coverage and contour finding algorithms, we suggest trajectory planning algorithms for single, two or four mobile robots. Various simulations reveal that the proposed algorithms improve time-efficiency in nuclear/chemical reconnaissance missions in the given area. Also we conduct basic experiments using a commercial mobile robot and verify the time efficiency of the proposed contour finding algorithms.

A neural network-based stochastic active contour model (NNS-SNAKE) for contour finding of distinct features

Contour finding of distinct features in 2-D/3-D images is essential for image analysis and computer vision. To overcome the potential problems associated with existing contour finding algorithms, we propose a framework, called the neural network-based stochastic active contour model (NNS-SNAKE), which integrates a neural network classifier for systematic knowledge building, an active contour model (also known as the "Snake") for automated contour finding using energy functions, and the Gibbs sampler to help the snake to find the most probable contour using a stochastic decision mechanism. Successful application of the NNS-SNAKE to extraction of several types of contours on magnetic resonance (MR) images is presented.

Periphere quantitative Computertomographie zur Knochenmineralmessung mit einem neuen speziellen QCT-Scanner

Some years ago we designed a special purpose gamma-ray QCT scanner for bone mineral measurements at the forearm. Currently, the first commercially built scanners (Stratec SCT 900) are being tested and clinically evaluated. The scanner uses an 125I photon source with a multidetector translate rotate fan beam geometry and newly developed semiconductor crystals. Quantification of trabecular bone and total bone at a standardized radial cross section gives an in-vivo reproducibility of less than 1%, which is superior compared to all other bone mineral measurement techniques. In-vitro reproducibility is between 0.2 and 0.4%, measured with calibration phantoms. The high in vivo reproducibility is ensured by automatic contour finding algorithms to determine the bone mass within a forearm cross sectional slice. A number of healthy men (n = 201) and women (n = 400) were examined to obtain a reference population for comparison to patients with generalized bone disease. We compared 182 osteoporotic females with healthy subjects. The high precision of the device and the high sensitivity of the measurement site considering changes of trabecular bone mass provide a very powerful means for monitoring bone mineral loss or changes due to therapy, as well as high diagnostic sensitivity.

Determination of object contour from projections for attenuation correction in cranial positron emission tomography.

A contour-finding algorithm is described by which the object periphery is outlined in positron emission tomography (PET) scans. The positions of maximum slopes are determined in the projections. These positions are shown to correspond well to the borders of the activity region. By backprojecting these points in the projections into the image plane, the size and shape of the region of activity must be outlined. Cord lengths across this region are then determined for attenuation correction. The accuracy of the algorithm has been tested on 10 patients with both PET and computed tomography (CT) examinations. Points on the periphery of the skull, as determined with the contour-finding algorithm in PET, coincided with those determined by CT, with a standard deviation of 1.5 mm and a maximum deviation of 3.5 mm. In testing the influence of the skull bone on the PET values, it was found that ignoring the higher attenuation coefficient of bone, 5 mm thick, in the attenuation correction caused an error in the determination of activity of 11% peripherally and 6% centrally in the object. These errors could be reduced to less than 1% by increasing the assumed attenuating region by 4.5 mm.