Candidate Areas Research Articles

Intelligent Segmentation and Measurement Model for Asphalt Road Cracks Based on Modified Mask R-CNN Algorithm

Nowadays, asphalt road has dominated highways around the world. Among various defects of asphalt road, cracks have been paid more attention, since cracks often cause major engineering and personnel safety incidents. Current manual crack inspection methods are time-consuming and labor-intensive, and most segmentation methods cannot detect cracks at the pixel level. This paper proposes an intelligent segmentation and measurement model based on the modified Mask R-CNN algorithm to automatically and accurately detect asphalt road cracks. The model proposed in this paper mainly includes a convolutional neural network (CNN), an optimized region proposal network (RPN), a region of interest (RoI) Align layer, a candidate area classification network and a Mask branch of fully convolutional network (FCN). The ratio and size of anchors in the RPN are adjusted to improve the accuracy and efficiency of segmentation. Soft non-maximum suppression (Soft-NMS) algorithm is developed to improve the segmentation accuracy. A dataset including 8,689 images (512 × 512 pixels) of asphalt cracks is established and the road crack is manually marked. Transfer learning is used to initialize the model parameters in the training process. To optimize the model training parameters, multiple comparison experiments are performed, and the test results show that the mean average precision (mAP) value and F1-score of the optimal trained model are 0.952 and 0.949. Subsequently, the robustness verification test and comparative test of the trained model are conducted and the topological features of the crack are extracted. Then, the damage area, length and average width of the crack are measured automatically and accurately at pixel level. More importantly, this paper develops an automatic crack detection platform for asphalt roads to automatically extract the number, area, length and average width of cracks, which can significantly improve the crack detection efficiency for the road maintenance industry.

A Face Detection Method Based on Skin Color Model and Improved AdaBoost Algorithm

This paper integrates skin color model and improved AdaBoost into a face detection method for high-resolution images with complex backgrounds. Firstly, the skin color areas were detected in a multi-color space. Each image was subject to adaptive brightness compensation, and converted into the YCbCr space, and a skin color model was established to solve face similarity. After eliminating the background interference by morphological method, the skin color areas were segmented to obtain the candidate face areas. Next, the inertia weight control factors and random search factor were introduced to optimize the global search ability of particle swarm optimization (PSO). The improved PSO was adopted to optimize the initial connection weights and output thresholds of the neural network. After that, a strong AdaBoost classifier was designed based on optimized weak BPNN classifiers, and the weight distribution strategy of AdaBoost was further improved. Finally, the improved AdaBoost was employed to detect the final face areas among the candidate areas. Simulation results show that our face detection method achieved high detection rate at a fast speed, and lowered false detection rate and missed detection rate.

Automatic brain extraction from 3D fetal MR image with deep learning-based multi-step framework.

Brain extraction is a fundamental prerequisite step in neuroimage analysis for fetus. Due to surrounding maternal tissues and unpredictable movement, brain extraction from fetal Magnetic Resonance (MR) images is a challenging task. In this paper, we propose a novel deep learning-based multi-step framework for brain extraction from 3D fetal MR images. In the first step, a global localization network is applied to estimate probability maps for brain candidates. Connected-component labeling algorithm is applied to eliminate small erroneous components and accurately locate the candidate brain area. In the second step, a local refinement network is implemented in the brain candidate area to obtain fine-grained probability maps. Final extraction results are derived by a fusion network with the two cascaded probability maps obtained from previous two steps. Experimental results demonstrate that our proposed method has superior performance compared with existing deep learning-based methods.

Letters of reference for PGY1 pharmacy residency candidates: A survey of residency program directors and opinion statement

AbstractIntroductionAlthough letters of reference (LORs) are required for postgraduate year one (PGY1) residency applications and used to distinguish between candidates, an evaluation of residency program director (RPD) perceptions and values surrounding LORs has not been described.ObjectiveThe aim of the study was to describe PGY1 RPDs' processes, values, and perceptions surrounding LORs.MethodsThis was a cross‐sectional descriptive survey. After assessing content validity, the Education and Training Practice and Research Network (PRN) taskforce sent a 25‐item electronic survey to 1266 PGY1 RPDs. Survey questions captured program demographics, LOR requirements, processes for LOR review, and perceptions about the value of LORs. Two open‐ended questions solicited suggestions for LORs and the existing standardized form; responses were evaluated via thematic analysis.ResultsA total of 291 (24%) programs completed the survey. LORs were rated as extremely or quite valuable by 82% of respondents and moderately valuable by 16% of respondents. Scoring rubrics for LOR evaluation were used by 79% of programs, and performance ratings were reported to impact interview consideration by a majority of programs. Accuracy, detailed comments, and inclusion of specific candidate strengths and areas for improvement were rated as extremely or quite important characteristics by 90% or more of respondents. Specific strengths were reported to be present in LORs more than half of the time by 81% of programs; however, accuracy, detailed comments, and inclusion of specific areas for improvement were only reported to be present about half the time or less by 41%, 63%, and 63% of respondents. Among 207 suggestions submitted to improve LOR utility, common themes included a desire for LOR authors to have candid discussions with candidates prior to letter writing, to contextualize their relationship with the candidate, and provide honest and specific assessments.ConclusionLORs remain highly valued by PGY1 RPDs during application review. Several notable findings were observed, resulting in recommendations for LOR authors, residency applicants, and the standardized LOR form.

A combined method of crater detection and recognition based on deep learning

The crater is one of the main obstacles that need to be avoided when Mars probe lands. In order to further improve the accuracy of crater detection, this paper proposes a combined detection method based on deep learning. Firstly, the random structured forest is trained offline to detect the edge information of craters. Secondly, according to the detected edge information of the crater, the candidate areas of the crater are determined with the morphological method. For the identified candidate areas of the crater, Alexnet network trained by transfer learning was used to identify crater areas. Compared with other methods, the proposed method has relatively good effect.

The Selection of Ornamental Plant for Landscape Design of Pollination Garden at Bogor Botanic Gardens

Bogor Botanic Gardens is one of the outstanding green areas in Bogor City. Thousands of plant species are growing and providing suitable habitats and feed for various pollinators, such as butterflies, bees, and wasps. A well-designed pollination garden will benefit the pollinators and also create an educational facility for the community. This study aimed to select and characterize species of flowering plants that effectively attract pollinators, particularly Lepidoptera and Hymenoptera, in the pollination garden candidate area. The observa-tion had been carried out on 12 ornamental plant species: Pennisetum purpureum, Cosmos sp., Pachystachys lutea, Cuphea hyssopifolia, Orthosiphon aristatus, Lantana camara, Citrus aurantifolia, Buddleja davidii, Ixora javanica, Nerium oleander, Hibiscus rosa-sinensis, and Aglaia odorata. The method for observing pollinators was a standard walking method. Pollinator photos, its number, its visit time, and the plants it visited were recorded. A total of 23 species of Lepidoptera (20 species from 4 families and 3 species were unidentified) and 9 species of Hymenoptera (4 families) were found in the area with the best obser-vation time at 08.30-09.30 (GMT+7). The most visited plants by pollinators were Cuphea hyssopifolia, Lantana camara and Orthosiphon aristatus. Lepidoptera and Hymenoptera preferred blue-purple and bright-orange-yellow flowers. In addition, the flower characters that might influence pollinator visits are also described. The result would be an essential recommendation for plant selection, landscape design, and construction of the pollination garden in Bogor Botanic Gardens.

Ship Detection in Multispectral Remote Sensing Images via Saliency Analysis

Despite the increasing visible optical remote sensing cameras equipped with panchromatic and four-band multispectral sensors, the application of multispectral data is still rarely used in the field of remote sensing ship target detection with the openly recognized challenge to improve detection precision. Towards this end, a ship target saliency detection method is proposed in the paper, which combines weighted least squares (WLS) with maximum symmetric surround (MSS), based on RGB-NIR four-band multispectral remote sensing images to locate the candidate area of ship targets quickly and accurately. The high frequency information of the NIR band image is extracted through the WLS filter and integrated into the RGB band image, and then the saliency of the image is analyzed. The detection results are combined with AIS to achieve complementary information for ship recognition. Some experiments show that the proposed method can effectively suppress the complex background information of clutter interference such as cloud waves and sea waves, highlight the ship target in low contrast scenes, and also increase recall and precision effectively.

Assessment of REDD+ Suitable Area for Sustainable Forest Management in Paraguay

This study extracted deforestation area and degraded forestland area, which are potential REDD+ (Reducing Emissions from Deforestation and Forest Degradation) project candidate areas in Paraguay using Land Cover Map (LCM) and Tree Cover Map (TCM). The REDD+ project objectives scenarios were set three stages: ‘afforestation and economic efficiency scenario’, ‘local capacity reinforcement scenario’, and ‘Infrastructure-oriented scenario’. And then, we evaluated the project unit suitable area of the REDD+ project. All scenarios selected the evaluation factors for each scenario in addition to the area ratio factors for deforestation area and degraded forestland area and weighted values were extracted by assigning category scores. As a result of the three scenarios comparison analysis, Concepcion state score was the highest. Within Concepcion state, the Belon district had the highest score, making it appropriate as a project unit REDD+ project candidate area in Paraguay, while the San Carlos district had the lowest score. This study can be used as basic data for selecting REDD+ project candidate area in Paraguay, and it is expected to contribute sufficiently to REDD+ project if additional data or information of social, cultural and economic sectors are secured.

Candidate site selection for nuclear power plants in Saudi Arabia using GIS

Many countries prefer to use renewable and clean energy sources over fossil fuels as fossil fuels are limited and have a negative impact on the environment. Although there are diverse, renewable, and clean sources of energy, such as solar, wind, and geothermal, the low and variable productivity of these resources impedes their application. Conversely, nuclear energy is a reliable, stable, and safe source of energy. The International Atomic Energy Agency recommends the use of geographic information systems in site planning of nuclear plants. Accordingly, this study seeks to apply the technical capabilities of geographic information systems in building a geographical database as well as data processing and analysis to support the decision making and examination of suitable sites for nuclear reactor facilities. In this study, a spatial suitability analysis methodology based on the multi-criteria decision-making method was used to identify candidate areas with suitable spatial characteristics for nuclear reactor facilities in the Kingdom of Saudi Arabia. Moreover, vector modeling was used to successfully determine potential sites over an area of ~ 4442 km2 on the eastern coast (Al-Ahsa, Al-Adaid Governorate) and over an area of ~ 600 km2 on the western coast (Umloj, Yanbu, Rabigh, Al-Laith, and Baysh) of the Kingdom.

Probability-based Mask R-CNN for pulmonary embolism detection

Pulmonary embolism (PE), a blockage of the lung artery, is common and sometimes fatal. Early diagnosis and treatment of PE can reduce the risk of associated morbidity and mortality. However, it is a huge challenge to accurately detect PE, particularly for the case of small segmental and subsegmental emboli. In this paper, a flexible probability-based Mask R-CNN model, namely P-Mask RCNN, is proposed for PE detection. Specifically, the feature map is firstly upsampled to enrich the local details of the small objects and to extract anchors at a higher density. Then, a candidate area is constructed based on the probability of the appearance of PE. Finally, we extract the anchors in the candidate area of the enlarged feature map for subsequent detection. Extracting anchors in the candidate area instead of the entire image can not only reduce both time and space consumption caused by the enlarging feature maps but also improve the detection performance by eliminating most invalid anchors. Compared with Mask R-CNN, the anchors extracted by the proposed P-Mask RCNN is closer to the ground truth. Extensive experimental results demonstrate the effectiveness and efficiency of the proposed approach. The source code of our method is available athttps://github.com/longkun-uestc/P_Mask_RCNN.

Optimal Dummy Pattern Design Method for PWB Warpage Control Using the Human-Based Genetic Algorithm.

In this work, a method that minimizes printed wiring board (PWB) warpage by dummy pattern design is proposed. This work suggests that dummy patterns are placed on a preset discretized location in the PWB to reduce the warpage. On each discretized candidate area, the dummy pattern can be set or unset. The warpage is numerically simulated based on direct modeling of the as-is PWB patterns to evaluate the warpage alongside the dummy pattern design set. The optimal pattern that minimizes warpage is determined using the human-based genetic algorithm where the objective function is evaluated by the structural simulation. The optimization method is realized in a spreadsheet that allows scripting language with which the input and output files of the simulation tool can be modified and read. Two different cases have been tested and the results show that the method can determine the optimal dummy patterns. The measured and simulated deflections agree well with each other. Moreover, it has been shown that certain dummy pattern designs that should reduce the warpage can be sought by the optimization.

A characteristic periglacial landform: Automated recognition and delineation of cryoplanation terraces in eastern Beringia

AbstractAutomated recognition and delineation of specific landforms and their constituent elements ranks among the most active areas of contemporary geomorphological research. This study contributes to that literature by applying semi‐ and fully automated recognition procedures to upland periglacial geomorphic landscapes. The Cryoplanation Terrace semi‐Automated Recognition (CTAR) algorithm utilizes basic terrain parameters to identify locations of cryoplanation terraces (CTs) from the high‐resolution ArcticDEM. Using a multistep process, candidate areas are identified based on morphometric characteristics. CTAR uses terrain derivatives to search ridges, hills, and mountains for flat areas bounded by abrupt breaks in slope. Because CTs are found exclusively in upland periglacial environments, some locations require that low‐lying areas be filtered out. To assess accuracy, CTAR was tested at five local study sites distributed across eastern Beringia, each containing multiple CTs delimited manually in a previous study. CTAR performed well, with an overall accuracy of 90%. A strong linear relationship exists between the size of CTAR‐delimited terraces and those identified in a previous study through air‐photo interpretation. In addition to identifying nearly all of the CTs in the five study areas, a fully automated version of the algorithm (GEE‐CTAR), implemented in Google Earth Engine, identified nearly 8,000 previously unmapped potential CTs in the Seward Peninsula region of western Alaska. The ability to identify CTs from digital elevation models provides a useful tool for recognizing and delineating upland periglacial topography. Objective recognition of large erosional landform elements created by periglacial processes is a critical step in developing the field of periglacial geomorphometry.

CLASSIFICATION OF RAILWAY ASSETS IN MOBILE MAPPING POINT CLOUDS

Abstract. Nowadays, mobile mapping systems are widely used to quickly collect reliable geospatial information of relatively large areas: thanks to such characteristics, the number of applications and fields exploiting their usage is continuously increasing. Among such possible applications, mobile mapping systems have been recently considered also by railway system managers to quickly produce and update a database of the geospatial features of such system, also called assets. Despite several vehicles, devices and acquisition methods can be considered for the data collection of the railway system, the predominant one is probably that based on the use of a mobile mapping system mounted on a train, which moves all along the railway tracks, enabling the 3D reproduction of the entire railway track area.Given the large amount of data collected by such mobile mapping, automatic procedures have to be used to speed up the process of extracting the spatial information of interest, i.e. assets positions and characteristics.This paper considers the problem of extracting such information for what concerns cantilever and portal masts, by exploiting a mixed approach. First, a set of candidate areas are extracted and pre-processed by considering certain of their geometric characteristics, mainly extracted by using eigenvalues of the covariance matrix of a point neighborhood. Then, a 3D modified Fisher vector-deep learning neural net is used to classify the candidates. Tests on such approach are conducted in two areas of the Italian railway system.

Joint deep separable convolution network and border regression reinforcement for object detection

The improvement of object detection performance mainly depends on the extraction of local information near the target area of interest, which is also the main reason for the lack of feature semantic information. Considering the importance of scene and semantic information for visual recognition, in this paper, the improvement of the object detection algorithm is realized from three parts. Firstly, the basic residual convolution module is fused with the separable convolution module to construct a depth-wise separable convolution network (D_SCNet-127 R-CNN). Then, the feature map is sent to the scene-level region proposal self-attention network to re-identify the candidate area. This part is composed of three parallel branches: semantic segmentation module, region proposal network, and region proposal self-attention module. Finally, this paper uses deep reinforcement learning combined with a border regression network to achieve precise location of the object, and improve the calculation speed of the entire model through a light-weight head network. This model can effectively solve the limitation of feature extraction in traditional object detection and obtain more comprehensive detailed features. The experimental on MSCOCO17, Pascal VOC07, and Cityscapes datasets shows that the proposed method has good validity and scalability.

Obstacle detection in a field environment based on a convolutional neural network security

ABSTRACT Information security has become an important subject in the artificial intelligence filed to handle big data. Most of the systems aim at obstacle detection on ordinary roads. In this paper, we proposed a method for detecting obstacles in a field environment based on convolutional neural network (CNN). Firstly, we propose a region of interest (ROI) extraction algorithm to deal with the suspected obstacle area. Secondly, we design a CNN model to classify the extracted feature maps of candidate areas. The experimental results indicate that the proposed method has high recognition accuracy and can detect obstacles effectively. Abbreviations: CNN: Convolutional Neural Network; ROI: Region of Interest; DBM: Deep Boltzmann Machines; AE: Auto-encoders; RPN: Region Proposal Network; ReLU: Rectified Linear Unit; RCNN: Regions with CNN Features; VGG-Net: Visual Geometry Group Net

Shadow detection in SAR images based on greyscale distribution, a saliency model, and geometrical matching

ABSTRACT Shadow information has been widely used in synthetic aperture radar interpretation, but corresponding shadow detection technology has not been given much attention in past studies. In this paper, we propose a hierarchical architecture based on greyscale distribution, a saliency model, and geometrical matching for shadow detection in SAR images. We find that the greyscale distribution of one Gaussian filtered image might contain a ‘distortion’ in its uphill part, which is the effect of shadow existence. Based on this distortion, a global threshold can be determined and then be used to segment candidate shadows. If there is no obvious distortion, a shadow saliency model is proposed as a substitute to extract such candidate areas. Usually, these candidate areas may contain some non-shadow components. According to the geometric relationships between shadow and object, we design a matching strategy to eliminate non-shadow parts from candidate regions. The remained areas are final shadow detection results. Experiments on two real datasets, Moving and Stationary Target Acquisition Recognition and MiniSAR, show that our method performs much better than two other published methods. The results demonstrate the effectiveness and feasibility of our proposed algorithm in practical SAR shadow detection tasks.

Detecting Pedestrians and Their Walk Directions Using a MY VISION System

This paper proposes a pedestrian detection method using a MY VISION system. The MY VISION system is an image processing system using an ego-camera which a user of the system possesses, and it is expected to be the third eye of those who are aged, visually impaired or even those who are absorbed in a mobile-phone while walking. In the proposed method, a flow area different from camera movement is extracted first on the image provided from a user’s head-mounted camera. The area is a candidate area where pedestrians may exist. Multiple scale cell-histograms of oriented gradients features are then calculated to detect pedestrians in the area. Histograms of optical flow feature is further computed to recognize the moving direction of the detected pedestrians. The proposed method was examined its performance experimentally and satisfactory results were obtained.

Target detection algorithm based on improved multi-scale SSD

The traditional SSD algorithm has a serious abstraction of feature extraction content, which makes it difficult to achieve effective detection of small targets. At the same time, the problem of feature layer fusion is difficult due to different scales. In this paper, an improved SSD based target detection algorithm is proposed. By introducing feature enhancement method, the adjustment steps of high-level feature size are omitted, Which makes it unnecessary to reduce the dimension of features, and at the same time, it uses the multi-scale candidate area which accords with the proportion of pedestrians in the detection network to enhance the feature extraction ability of small targets, effectively improves the accuracy and operation speed of SSD algorithm, and saves the loss of the network.

Urban expansion simulation from the perspective of land acquisition-based on bargaining model and ant colony optimization

As a major method to serve the demand for land requirement in Chinese urban construction, land acquisition has been intensified since the speeding up of urbanization. However, clashes arise during the process of land acquisition, out of conflicts of interests, which to some extent have affected the development and direction of urban expansion, as well as social harmony and stability. Therefore, the urban expansion simulation should be based on smooth land acquisition technique. In this study, urban expansion was simulated from the perspective of land acquisition, based on coupled bargaining model and modified ant colony optimization (ACO) algorithm. First, the bargaining model with fairness preferences should be set up for the government and farmers, to search for the candidate areas for urban expansion, both sides of which could reach a consensus without conflicts of interest. Then, urban expansion simulation is applied to all the candidate areas, taking advantage of the modified ACO algorithm. The model considered the built-up area in Wuhan city as the demonstration area and simulated conditions of land use in 2016 and 2026. The result showed that the coupled model could simulate decision-making behavior of the government and farmers in land acquisition veritably, so as to protect farmers' economic interests, with an increase of over 50% on average, and ensure government's appropriate profit. Moreover, the simulation accuracies of the coupled model was found to be better than that of the traditional cellular automata model, and the Kappa coefficient was 0.65, which supports the effectiveness of the model in simulating urban expansion. Further, it was estimated that the urban land use of Wuhan will cover 516.22 km2 in 2026, and the southeastern part of the city will be the hot spot area of urban expansion.

Automated fibroglandular tissue segmentation in breast MRI using generative adversarial networks

Fibroglandular tissue (FGT) segmentation is a crucial step for quantitative analysis of background parenchymal enhancement (BPE) in magnetic resonance imaging (MRI), which is useful for breast cancer risk assessment. In this study, we develop an automated deep learning method based on a generative adversarial network (GAN) to identify the FGT region in MRI volumes and evaluate its impact on a specific clinical application. The GAN consists of an improved U-Net as a generator to generate FGT candidate areas and a patch deep convolutional neural network (DCNN) as a discriminator to evaluate the authenticity of the synthetic FGT region. The proposed method has two improvements compared to the classical U-Net: (1) the improved U-Net is designed to extract more features of the FGT region for a more accurate description of the FGT region; (2) a patch DCNN is designed for discriminating the authenticity of the FGT region generated by the improved U-Net, which makes the segmentation result more stable and accurate. A dataset of 100 three-dimensional (3D) bilateral breast MRI scans from 100 patients (aged 22–78 years) was used in this study with Institutional Review Board (IRB) approval. 3D hand-segmented FGT areas for all breasts were provided as a reference standard. Five-fold cross-validation was used in training and testing of the models. The Dice similarity coefficient (DSC) and Jaccard index (JI) values were evaluated to measure the segmentation accuracy. The previous method using classical U-Net was used as a baseline in this study. In the five partitions of the cross-validation set, the GAN achieved DSC and JI values of 87.0 ± 7.0% and 77.6 ± 10.1%, respectively, while the corresponding values obtained through by the baseline method were 81.1 ± 8.7% and 69.0 ± 11.3%, respectively. The proposed method is significantly superior to the previous method using U-Net. The FGT segmentation impacted the BPE quantification application in the following manner: the correlation coefficients between the quantified BPE value and BI-RADS BPE categories provided by the radiologist were 0.46 ± 0.15 (best: 0.63) based on GAN segmented FGT areas, while the corresponding correlation coefficients were 0.41 ± 0.16 (best: 0.60) based on baseline U-Net segmented FGT areas. BPE can be quantified better using the FGT areas segmented by the proposed GAN model than using the FGT areas segmented by the baseline U-Net.