Contour Points Research Articles

Single Person Identification and Activity Estimation in a Room from Waist-Level Contours Captured by 2D Light Detection and Ranging.

To develop socially assistive robots for monitoring older adults at home, a sensor is required to identify residents and capture activities within the room without violating privacy. We focused on 2D Light Detection and Ranging (2D-LIDAR) capable of robustly measuring human contours in a room. While horizontal 2D contour data can provide human location, identifying humans and activities from these contours is challenging. To address this issue, we developed novel methods using deep learning techniques. This paper proposes methods for person identification and activity estimation in a room using contour point clouds captured by a single 2D-LIDAR at hip height. In this approach, human contours were extracted from 2D-LIDAR data using density-based spatial clustering of applications with noise. Subsequently, the person and activity within a 10-s interval were estimated employing deep learning techniques. Two deep learning models, namely Long Short-Term Memory (LSTM) and image classification (VGG16), were compared. In the experiment, a total of 120 min of walking data and 100 min of additional activities (door opening, sitting, and standing) were collected from four participants. The LSTM-based and VGG16-based methods achieved accuracies of 65.3% and 89.7%, respectively, for person identification among the four individuals. Furthermore, these methods demonstrated accuracies of 94.2% and 97.9%, respectively, for the estimation of the four activities. Despite the 2D-LIDAR point clouds at hip height containing small features related to gait, the results indicate that the VGG16-based method has the capability to identify individuals and accurately estimate their activities.

Underwater terrain-matching algorithm based on improved iterative closest contour point algorithm

Although an autonomous underwater vehicle (AUV) is noted for its good autonomy, concealment and anti-interference ability, errors in its inertial navigation system (INS) inevitably increase over time, leading to positional failure during long-term voyages. Terrain-assisted navigation can help the INS to correct its position. The traditional iterative closest contour point (ICCP) achieves high matching accuracy when the initial position error of the INS is small, but is prone to mismatching when the initial error is large. This study combines ICCP with particle swarm optimization (PSO) to overcome this problem. First, the global optimization ability of PSO is improved by changing the acceleration factor and introducing an artificial bee colony (ABC) onlooker bee greedy search (ABC- ωAPSO). Second, the Euclidean distance of ICCP is replaced by the Mahalanobis distance to abate the influence of system error on the matching accuracy. Finally, the initial position error is reduced by rough matching using the ABC- ωAPSO, which has global optimization capability. Next, fine matching is performed by ICCP. This two-step process resolves the sensitivity problem of ICCP to the initial position error. The experimental results revealed a good matching effect after the double-matching procedure. When the initial INS errors were 0.55′ to the east and 0.55′ to the north, the matching error was reduced to 89.3 m, suggesting that the approach can realize autonomous passive navigation of AUVs.

A virtual phantom for patient-specific QA On A 1.5T MR-linac.

Create a virtual ArcCHECK-MR phantom, customized for a 1.5TMR-linac, with consideration of the different density regions within the quality assurance (QA) phantom, aiming to streamline the utilization of this specialized QA device. A virtual phantom was constructed in the treatment planning system (TPS) to replicate the ArcCHECK-MR's composition, consisting of five distinct layers: "Outer" (representing the outer PMMA ring), "Complex" (simulating the printed circuit boards), "Detectors" (encompassing the detector area), "Inner" (signifying the inner PMMA ring) and "Insert" (representing the PMMA insert). These layers were defined based on geometric data and represented as contour points on a set of dummy CT images. Additionally, a setup platform was integrated as contoured structures. To determine the relative electron density (RED) values of the external and internal PMMA components, measurements were taken at 25 points in the insert using an ion chamber. A novel method for establishing the exit/entrance dose ratio (EEDR) for ArcCHECK-MR was introduced. The RED of higher density region was derived by evaluating the local gamma index passing rate results with criteria of 2% dose difference and 2mm distance-to-agreement. The performance of the virtual phantom was assessed for Unity 7 FFF beams with a 1.5T magnetic field. The radii of the five ring structures within the virtual phantom measured 133.0mm, 110.0mm, 103.4mm, 100.0mm, and 75.0mm for the "Outer," "Complex," "Detectors," "Inner" and "Insert" regions, respectively. The RED values were as follows: ArcCHECK-MR PMMA had a RED of 1.130, "Detectors" were assumed to have a RED of 1.000, "Complex" had a RED of 1.200, and the setup QA phantom justified a RED of 1.350. Early validation results demonstrate that the 5-layer virtual phantom, when compared to the commonly used bulk overridden phantom, offers improved capability in MR-linac environments. This enhancement led to an increase in passing rates for the local gamma index by approximately 5 ∼ 6%, when applying the criteria of 2%, 2mm. We have successfully generated a virtual representation of the distinct regions within the ArcCHECK-MR using a TPS, addressing the challenges associated with its use in conjunction with a 1.5T MR-linac. We consistently observed favorable local gamma index passing rates across two 1.5T MR-linac and ArcCHECK-MR unit combinations. This approach has the potential to minimize uncertainties in the creation of the QA phantom for ArcCHECK-MR across various institutions.

SatellStitch: Satellite Imagery-Assisted UAV Image Seamless Stitching for Emergency Response without GCP and GNSS

Rapidly stitching unmanned aerial vehicle (UAV) imagery to produce high-resolution fast-stitch maps is key to UAV emergency mapping. However, common problems such as gaps and ghosting in image stitching remain challenging and directly affect the visual interpretation value of the imagery product. Inspired by the data characteristics of high-precision satellite images with rich access and geographic coordinates, a seamless stitching method is proposed for emergency response without the support of ground control points (CGPs) and global navigation satellite systems (GNSS). This method aims to eliminate stitching traces and solve the problem of stitching error accumulation. Firstly, satellite images are introduced to support image alignment and geographic coordinate acquisition simultaneously using matching relationships. Then a dynamic contour point set is constructed to locate the stitching region and adaptively extract the fused region of interest (FROI). Finally, the gradient weight cost map of the FROI image is computed and the Laplacian pyramid fusion rule is improved to achieve seamless production of the fast-stitch image map with geolocation information. Experimental results indicate that the method is well adapted to two representative sets of UAV images. Compared with the Laplacian pyramid fusion algorithm, the peak signal-to-noise ratio (PSNR) of the image stitching results can be improved by 31.73% on average, and the mutual information (MI) can be improved by 19.98% on average. With no reliance on CGPs or GNSS support, fast-stitch image maps are more robust in harsh environments, making them ideal for emergency mapping and security applications.

Extraction of Building Roof Contours from Airborne LiDAR Point Clouds Based on Multidirectional Bands

Because of the complex structure and different shapes of building contours, the uneven density distribution of airborne LiDAR point clouds, and occlusion, existing building contour extraction algorithms are subject to such problems as poor robustness, difficulty with setting parameters, and low extraction efficiency. To solve these problems, a building contour extraction algorithm based on multidirectional bands was proposed in this study. Firstly, the point clouds were divided into bands with the same width in one direction, the points within each band were vertically projected on the central axis in the band, the two projection points with the farthest distance were determined, and their corresponding original points were regarded as the roof contour points; given that the contour points obtained based on single-direction bands were sparse and discontinuous, different banding directions were selected to repeat the above contour point marking process, and the contour points extracted from the different banding directions were integrated as the initial contour points. Then, the initial contour points were sorted and connected according to the principle of joining the nearest points in the forward direction, and the edges with lengths greater than a given threshold were recognized as long edges, which remained to be further densified. Finally, each long edge was densified by selecting the noninitial contour point closest to the midpoint of the long edge, and the densification process was repeated for the updated long edge. In the end, a building roof contour line with complete details and topological relationships was obtained. In this study, three point cloud datasets of representative building roofs were chosen for experiments. The results show that the proposed algorithm can extract high-quality outer contours from point clouds with various boundary structures, accompanied by strong robustness for point clouds differing in density and density change. Moreover, the proposed algorithm is characterized by easily setting parameters and high efficiency for extracting outer contours. Specific to the experimental data selected for this study, the PoLiS values in the outer contour extraction results were always smaller than 0.2 m, and the RAE values were smaller than 7%. Hence, the proposed algorithm can provide high-precision outer contour information on buildings for applications such as 3D building model reconstruction.

Application of 3D Printing Technology in Customized Garment Production

Abstract In this paper, human body images are acquired by DSLR camera shooting, and the photos are grayscaled based on the HIS model. Then, the different operations weaken the similar parts between the images, and the other parts of the photos are highlighted. The image is separated from the background by image binarization. After the contour detection of the image, the pixel coordinates of the contour points in the image are deposited into a dynamic array by the size judgment of the pixel values. The subsequent size calculation is called to make a pad. According to the human body, clothing size automation measurement has been derived from the human body characteristics of the size, the use of three-dimensional printing technology to produce clothing to meet consumer demand, and three-dimensional printing technology to support the production of customized clothing under the experimental analysis. It is verified that the data difference between the actual size of the subject’s ready-made clothes and the size of the clothes produced based on 3D printing technology is kept within ±0.48cm. The virtual fitting pressure is less than the comfort threshold of women’s upper body pressure (2.75kpa), which indicates that the clothes produced by this research will not cause physiological discomfort to the human body and, at the same time, meet the consumer’s personalized clothing customization needs.

단순 방사선 영상기반 바이오마커를 활용한 흉·요추의 추나의학적 변위 진단 방법

Objectives This study aimed to propose biomarkers for diagnosing Chuna manual therapy (CMT) based on X-ray images in the thoracic and lumbar spines.
 Methods Through a literature review and expert consensus process, diagnostic bio- markers for CMT were selected based on the listing system in thoracic and lumbar ra- diograph anterior-posterior (AP) and lateral views.
 Results 1. Diagnostic biomarkers were derived from four points on the outer contour of the vertebral body in the thoracic and lumbar spine radiograph lateral view, en- abling the diagnosis of flexion and extension malposition. 2. Additional diagnostic bi- omarkers were identified in the thoracic and lumbar radiographAP view, utilizing points on the outer contour of the vertebral body. These biomarkers facilitate the di- agnosis of lateral bending. Moreover, biomarkers derived from the innermost point of the pedicle contour allow for the diagnosis of rotation malposition. 3. Furthermore, through the biomarkers proposed in this study, all malpositions of the thoracolumbar spines and complex Type I and II malpositions can be diagnosed in CMT.
 Conclusions The biomarkers reported in this study consist of minimal points to de- termine the position of the vertebral body, providing the advantage of simplicity while minimizing potential errors during the CMT diagnostic process. Further clinical re- search and the development of related programs should be pursued to expand the evidence for CMT.

Analysis of the stress state at the contours of underground horizontal workings of elliptical cross-section shape subjected to all-round uniform pressure

The results of the research to determine the stress state on the contours of underground horizontal excavations, the cross-section of which has the form of an ellipse of any size and shape are presented. When changing the intensity of tensile uniform pressure applied to the points of elliptical shape cross-section excavation contours, the given values of excavation depths, ellipse parameters and rock side spreading coefficient values, the areas of the contours with tensile and compressive stresses are identified. If the depths of workings of the considered form of cross-section change, the areas where tensile and compressive stresses act at given values of the intensity of tensile uniform pressure applied to the points of working's contours, ellipse parameters and values of the coefficient of rock lateral spreading are highlighted. The graphical images of the normal tangential stresses, acting on the contours of the considered workings are given. The ranges of intensity of the tensile uniform pressure, at which, depending on the depth of excavation, values of ellipse parameters and the coefficient of rock lateral spread, tensile and (or) compressive stresses act in the points of excavation contours have been determined. The given relations and results can be applied to calculate the values of the permissible values of all-round uniform pressure in the points of the workings contours and to determine the permissible depths of their embedding. As a criterion for determining the values of these values, the condition of absente on the contours of excavations of points in which the normal tangential stresses exceed the limits of tensile and compression strength of the rock can be accepted.

Perturbation of singular integral operators with piecewise continuous coefficients

In the paper it is shown that the property of singular integral operators with piecewise continuous coefficients to be Noetherian is stable with respect to their perturbation with certain non-compact operators. An example is constructed showing that the corner points of the integration contour significantly affect the Noetherian property of singular operators with translations. These results are obtained using the symbol of the singular operators on contours with corner points, symbol, which is also determined.

Perturbation of singular integral operators with piecewise continuous coefficients

In the paper it is shown that the property of singular integral operators with piecewise continuous coefficients to be Noetherian is stable with respect to their perturbation with certain non-compact operators. An example is constructed showing that the corner points of the integration contour significantly affect the Noetherian property of singular operators with translations. These results are obtained using the symbol of the singular operators on contours with corner points, symbol, which is also determined.

Joint object contour points and semantics for instance segmentation

AbstractThe edges of objects are of great significance to the task of instance segmentation. However, most of the current popular deep neural networks do not pay much attention to the object edge information. More importantly, using the down‐sampling pooling layer in the deep learning network, the edge detail information of the object will be lost. To address this issue, inspired by the manual annotation process, we propose Mask Point R‐CNN aiming at promoting the neural network's attention to the object boundary. Specifically, we introduce the auxiliary task of object contour point detection on the Mask R‐CNN framework, which can effectively improve the gradient flow between different tasks by multi‐task learning and repairing objects' boundary information via feature fusion. Consequently, the model can be more sensitive to the edges of the object and capture more geometric features. Quantitatively, the experimental results show that our Mask Point R‐CNN outperforms vanilla Mask R‐CNN by 3.8% on the Cityscapes dataset and 0.8% on the COCO dataset.

A system and method for person identification and positioning incorporating object edge detection and scale-invariant feature transformation

During the overhead crane, due to the complex background and large image size, the direct processing algorithm usually leads to time-consuming operation and difficult to achieve accurate personnel positioning. To address these issues, we propose the feature points recognition and matching algorithm incorporating object edge detection and scale-invariant feature transform (SIFT), which is applied to the designed object detection and positioning algorithm for binocular stereo vision. The proposed method applies the Canny operator to extract contours from the locally cropped regions obtained using YOLOv5. The contour points are used as the key points of SIFT, and corresponding feature points are created descriptors, and feature points matching are performed. Experimental results indicate that the proposed method can filter feature points successfully. Moreover, the binocular stereo vision system with the binocular positioning algorithm designed has good performance, with the overall depth direction positioning error within 5% in a range of 12 m.

An automatic reduction method of 3D bone fragments based on a novel section contour point descriptor.

Comminuted fractures are orthopedic traumas with greater surgical difficulty. In clinical treatment, a great challenge is precise reduction of multiple broken bone fragments; Another great challenge is personalized and precise internal fixation after reduction. For these two issues, we designed an automated method framework for precise reduction and internal fixation of comminuted fractures. First, the Gaussian mixture model (GMM) is used to distinguish section points and noise points in a broken bone model; Second, ellipse fitting is carried out to achieve section points matching and a descriptor is proposed to describe the section features; Then, the Convolution Auto-Encoder (CAE) and genetic algorithm are used to extract feature vectors; Finally, after broken bone models registration, internal fixed plate can be reconstructed. Three verification experiments for comminuted bone fracture show this method has high accuracy and good efficiency. It can provide support for minimally invasive treatment for comminuted fractures.

Buckling performance of inner corrugated pressure shells under external hydrostatic pressure

The most fundamental and central component of submarines and deep-sea space stations is the cylindrical pressure shell. In this paper, a pressure shell with inter corrugated structure was proposed to significantly improve its compression performance. 7 pressure shells including 6 inner corrugated pressure shells and 1 cylindrical pressure shell were fabricated and hydrostatic pressure tested. Geometrical deviations including thickness deviation and out-of-roundness were considered between the actual fabricated pressure shells and the ideal models. 3D scanner was used to measure the 3D contour point cloud data of all shells and the reverse modeling was performed. Nonlinear finite element analysis of reverse modeled shells containing actual geometric imperfections were performed to predict their buckling performance under external hydrostatic pressure. The results show that the corrugated structure can significantly increase the critical buckling load of the pressure shell. The failure positions of most pressure shells in the critical buckling state were mainly located near the middle region of the shell. The numerical calculation results were in good agreement with the experimental results. This research can provide good guidance for the design of pressure structures for large deep-sea equipment such as submarines and deep-sea space station.

Checkerboard corner detection method based onneighborhood linear fitting.

To improve the calibration accuracy of a vision measurement system, a checkerboard corner detection method based on linear fitting of the checkerboard local contour is proposed. First, by binarization and morphological dilation of the checkerboard image, the coordinates of two adjacent vertices of adjacent dark squares are obtained; the midpoint of the two vertices is taken as the reference point; the reference dotted array is obtained; and the Zernike moment subpixel method is used to obtain the checkerboard contour data points in the neighborhood of each reference point. Finally, the contour points are classified according to the orientation based on the reference points; two intersecting lines are fitted; and the intersection of the two lines is exactly the corner point that we want to find. A camera calibration experiment was conducted on the same group of checkerboard images. The results show that the calibration results of the corner points obtained based on this method are highly consistent with the OpenCV library function method and the MATLAB Toolbox calibration method, and the reprojection error is smaller. At the same time, it is robust to changes in the light source brightness.

The real-time vision measurement of multi geometric information of the rolled ring in the hot ring rolling process based on deep learning

Hot ring rolling (HRR) is an advanced incremental metal-forming technology to manufacture high-performance rings. Nowadays, the mechanical or laser-sensors measurement methods can only obtain one value of the dimensions on the ring circumference per time unit, the measurement data is single and the information is limited. This paper presents a real-time vision measurement method for multi geometric information of the rolled ring such as ring’s diameter, center position, circularity and growth speed in HRR process. Firstly, the deep learning model is constructed to intelligently and quickly identify rolled ring targets from complex image backgrounds, then the adaptive linear gray scale transformation algorithm is proposed to adjust the gray scale contrast of the image according to changes in ring temperature, the interference of oxide scales and flying chips are eliminated by morphology operation. Finally, the least square method is used to fit the contour points, and further the rolled ring’s geometric information and corresponding variation trend are calculated. Several experiments are conducted on a vertical HRR mill, the geometric information of the rolled ring during the HRR is measured in real-time under the interference of harsh working conditions. For the ring with outer diameter of 350 mm, the measurement error of ring’s outer and inner diameter is less than 0.8 mm, the average processing time per image takes about 80 ms. The research provides a basis for accurate measurement of HRR process.

Sphericity measurement based on telecentric imaging mechanism with image distortion correction and eccentricity error compensation

Sphericity is an essential geometric parameter for determining the processing manufacturing quality of spherical components. This study presents a sphericity measurement method based on telecentric imaging mechanism with image distortion correction and eccentricity error compensation. A sphericity measurement system is designed, and the projection images of the cross-section on the equatorial plane of the measured sphere are acquired using a line scan camera with bilateral telecentric lenses. An analytical camera model for this sphericity measurement system is established by considering major sources of image distortion. The subpixel edge points of the projection images are extracted and converted to cross-sectional profiles on the equatorial plane according to the camera model. An eccentricity error compensation strategy for contour matching of the obtained cross-sectional profile on the equatorial plane is also developed to obtain the three-dimensional contour point coordinates of the sphere. Sphericity is then calculated according to the least square criterion. The effectiveness of the proposed method is verified by measuring two steel balls with diameter in 20 mm and 30 mm. The sphericity measurement error is reduced by 0.3383 mm and 0.4653 mm respectively after image distortion correction and is reduced by 0.2268 mm and 0.0795 mm respectively after eccentricity error compensation. The proposed method can be extended to form error measurements for opaque components with rotary structure on the outer contour.

End-PolarT: Polar Representation for End-to-End Scene Text Detection

Deep learning has achieved great success in text detection, where recent methods adopt inspirations from segmentation to detect scene texts. However, most segmentation based methods have high computation cost in pixel-level classification and post refinements. Moreover, they still faces challenges like arbitrary directions, curved texts, illumination and so on. Aim to improve detection accuracy and computation cost, we propose an end-to-end and single-stage method named as End-PolarT network by generating contour points in polar coordinates for text detection. End-PolarT not only regress locations of contour points instead of pixels to relieve high computation cost, but also fits with intrinsic characteristics of text instances by centers and contours to suppress mislabeling boundary pixels. To cope with polar representation, we further propose polar IoU and centerness as key parts of loss functions to generate effective paradigms for text detection. Compared with the existing methods, End-PolarT achieves superior results by testing on several public datasets, thus keeping balance between efficiency and effectiveness in complicated scenes.

Body weight estimation of beef cattle with 3D deep learning model: PointNet++

It is necessary to measure the live body weight (BW) of beef cattle to assess their health and nutritional degree. Instead of continual manual measuring in holding frames or crates currently used in the breeding industry, this paper has proposed a novel non-contact BW prediction with 3D deep learning model PointNet++. The three-dimensional (3D) point cloud data (PCD) of the whole contour surface of a beef cattle is proximally obtained by a light detection and ranging (LiDAR) sensor. To preserve the quality of the silhouette, the fusion of classical PCD preprocessing is employed. To better the accuracy of manual selection to the position of body conformation, the data extension of PCD and dataset annotations expressed by the feature regions are employed. The PCD part segmentation of deep learning models is validated and redefined with datasets of beef cattle, where the PointNet++ model with the highest accuracy of 95.1% is finally redesigned. With the advantages of being suitable for most beef breeds for the classical Johnson’s live BW prediction, the Gaussian curvature and principal curvature of contour points of feature regions are calculated to extract the body measurement points of body size, and the body length (BL) and chest girth (CG) are calculated and fit-calibrated. With the experimentation of 100 cattle in breeding farms, the results of live BW prediction have shown that the mean absolute percentage error (MAPE) is 3.2%, the root mean square error (RMSE) is 10.2 kg. The effective and practical resolution of BW estimation can improve the weighing process and animal welfare in large-scale breeding farms, and the real silhouette dataset of large animals can help the automated recognition and classification of body conformation.

Construction of a physical-mathematical model of oscillations of the unbalanced vibrator of the pneumatic sorting table

A physical-mathematical model of oscillations of unbalanced vibrators of a pneumatic sorting table as non-stationary oscillations of an impulse-loaded round plate with various options for fixing its contour has been built. The axisymmetric non-stationary oscillations of a round plate supported by a one-sided round base were considered in two ways of fixing its contour, namely, when it is tightly clamped and freely supported. It was assumed that the linearly elastic base resists only compression and does not perceive stretching. It is shown that for certain durations of the transverse force pulse in time, the amplitude of the deflection of the middle of the plate in the direction of action of the external pulse can be smaller than the amplitude of the deflection in the opposite direction. At the same time, in the second case, there is no contact of the plate with the base. It has been proven that this dynamic effect of the asymmetry of the elastic characteristic of the system also applies to bending moments and is more clearly manifested when the contour is freely supported than when it is tightly clamped. For rectangular and sinusoidal pulses, closed-loop solutions of the equations of motion of the plate during its contact with the base and after separation from the base were constructed. Compact formulas were derived for calculating the amplitudes of positive and negative deflections in both directions from the zero position of static equilibrium. Formulas have been obtained for calculating the time it takes for the plate to obtain extreme deflection values, which is achieved due to the selection of a special axisymmetric distribution of dynamic pressure on the plate. Under such a load, the plate simultaneously detaches from the base at all points, except for the contour points, which reduces the nonlinear boundary value problem to a sequence of two linear problems. Numerical integration of the differential equation was carried out to check the reliability of the constructed analytical solutions. Adequacy of the model was proven for the following values of initial parameters: modulus of elasticity, 2.1·1011 Pa; the Poisson ratio of the plate material, 0.25; plate thickness, 7...10 mm; the maximum pressure on the plate, 4·103 Pa; the bending stiffness of the plate, 6402.6667 N·m