Distance Transform Research Articles

HyBoDT: Hybrid Bounded Distance Transforms of Trimmed NURBS Models

Abstract Distance field representation of objects in 3D space has several applications such as shape manipulation, graphics rendering, path planning, etc. Distance transforms (DTs) are discrete representations of distance fields in a regular voxel grid. The two main limitations of using distance transforms are that they are compute-intensive, and there are errors introduced while representing the object using DTs. In this work, we develop a hybrid graphics processing unit (GPU)-accelerated marching wavefront method for computing DTs of models composed of trimmed non-uniform rational B-splines (NURBS) surfaces with theoretical bounds. Our hybrid marching approach eliminates the error due to calculating approximate distances by marching. We also calculate the bounds on the error introduced due to the tessellation of the trimmed NURBS surfaces and calculate the propagation of these bounds in computing the DT. Finally, we present computation times for both 2D and 3D GPU DTs of test objects. We show that our GPU-accelerated approach is significantly faster than existing CPU-based methods.

Facial Expression Recognition Using Weighted Distance Transform

Facial emotions of human transfer non-verbal signal, which have a dynamic role in interactive communication. Human machine interface evolves according to facial expression recognition because both have a significant relationship. Psychology, ethical science and robotics are necessary applications of facial expression recognition. There have been a lot of work already done on feature extraction, face detection and the famous techniques used for expression recognition. Weighted distance is the basic method of this research. It is used for recognition of all basic human emotions, angry, happy, disgust, fear, neutral, sad and surprise. For the extraction of weighted distance paths, the fast marching algorithm is used and seed point has been taken on the nose tip of the human face. Diverse number of paths also has been taken, and it has an effect on facial expression recognition. Intensity variation is the main motivation to use Weighted Distance Transform. Twenty points are labelled for the calculation of feature vectors. Different mathematical measures are calculated as a feature vector of this geometric representation. In classification KNN is used and it illustrates reasonable result. In the end validation is done with famous techniques of facial expression recognition.

DINs: Deep Interactive Networks for Neurofibroma Segmentation in Neurofibromatosis Type 1 on Whole-Body MRI.

Neurofibromatosis type 1 (NF1) is an autosomal dominant tumor predisposition syndrome that involves the central and peripheral nervous systems. Accurate detection and segmentation of neurofibromas are essential for assessing tumor burden and longitudinal tumor size changes. Automatic convolutional neural networks (CNNs) are sensitive and vulnerable as tumors' variable anatomical location and heterogeneous appearance on MRI. In this study, wepropose deep interactive networks (DINs) to address the above limitations. User interactions guide the model to recognize complicated tumors and quickly adapt to heterogeneous tumors. We introduce a simple but effective Exponential Distance Transform (ExpDT) that converts user interactions into guide maps regarded as the spatial and appearance prior. Comparing with popular Euclidean and geodesic distances, ExpDT is more robust to various image sizes, which reserves the distribution of interactive inputs. Furthermore, to enhance the tumor-related features, we design a deep interactive module to propagate the guides into deeper layers. We train and evaluate DINs on three MRI data sets from NF1 patients. The experiment results yield significant improvements of 44% and 14% in DSC comparing with automated and other interactive methods, respectively. We also experimentally demonstrate the efficiency of DINs in reducing user burden when comparing with conventional interactive methods.

Extracting Buildings from Remote Sensing Images Using a Multitask Encoder-Decoder Network with Boundary Refinement

Extracting buildings from high-resolution remote sensing images is essential for many geospatial applications, such as building change detection, urban planning, and disaster emergency assessment. Due to the diversity of geometric shapes and the blurring of boundaries among buildings, it is still a challenging task to accurately generate building footprints from the complex scenes of remote sensing images. The rapid development of convolutional neural networks is presenting both new opportunities and challenges with respect to the extraction of buildings from high-resolution remote sensing images. To capture multilevel contextual information, most deep learning methods extract buildings by integrating multilevel features. However, the differential responses between such multilevel features are often ignored, leading to blurred contours in the extraction results. In this study, we propose an end-to-end multitask building extraction method to address these issues; this approach utilizes the rich contextual features of remote sensing images to assist with building segmentation while ensuring that the shape of the extraction results is preserved. By combining boundary classification and boundary distance regression, clear contour and distance transformation maps are generated to further improve the accuracy of building extraction. Subsequently, multiple refinement modules are used to refine each part of the network to minimize the loss of image feature information. Experimental comparisons conducted on the SpaceNet and Massachusetts building datasets show that the proposed method outperforms other deep learning methods in terms of building extraction results.

Deep active contours using locally controlled distance vector flow

Active contours model (ACM) has been extensively used in computer vision and image processing. In recent studies, convolutional neural networks (CNNs) have been combined with ACM replacing the user in the process of contour evolution and image segmentation to eliminate limitations associated with ACM dependence on energy functional parameters and initialization. However, prior studies did not aim for automatic initialization, which is addressed in this article. In addition to manual initialization, current methods are highly sensitive to the initial location and fail to delineate borders accurately. We propose a fully automatic image segmentation method to address problems of manual initialization, insufficient capture range, and poor convergence to boundaries, in addition to the problem of assignment of energy functional parameters. We train two CNNs, one of which generating ACM weighting parameters and the other generating a ground truth mask to extract distance transform (DT) and an initialization circle. DT is used to form a vector field pointing from each pixel of the image towards the closest ground truth boundary point. Vector magnitudes are equal to the Euclidean distance between each pixel and the closest ground truth boundary point. We evaluate our method on four publicly available datasets, including two building instance segmentation datasets, i.e., Vaihingen and Bing huts, and two mammography image datasets, INBreast and DDSM-BCRP. Our approach achieves state-of-the-art results in mean Intersection over Union (mIoU), Dice similarity coefficient and Boundary F-score (BoundF) with the values of 92.33%, 92.44%, and 86.57% for Vaihingen dataset, and 87.12%, 86.86%, and 66.91% for Bing huts dataset. We obtained the Dice similarity coefficient values of 94.23% and 90.89% for the INBreast and DDSM-BCRP, respectively.

Robust Hand Gesture Recognition Based on RGB-D Data for Natural Human–Computer Interaction

To naturally interact with virtual environment by hand gesture, this paper presents a robust RGB-D data based recognition method of static and dynamic hand gesture. Firstly, for static hand gesture recognition, starting from the hand gesture contour extraction, the palm center is identified by Distance Transform (DT) algorithm. The fingertips are localized by employing the K-Curvature-Convex Defects Detection algorithm (K-CCD). On the basis, the distances of the pixels on hand gesture contour to palm center and the angle between fingertips are considered as the auxiliary features to construct a multimodal feature vector, and then recognition algorithm is presented to robustly recognize the static hand gestures. Secondly, combining Euclidean distance between hand joints and shoulder center joint with the modulus ratios of skeleton features, this paper generates a unifying feature descriptor for each dynamic hand gesture and proposes an improved dynamic time warping (IDTW) algorithm to obtain recognition results of dynamic hand gestures. Finally, we conduct extensive experiments to test and verify the static and dynamic hand gesture recognition algorithm and realize a low-cost real-time application of natural interaction with virtual environment by hand gestures.

A framework for neighbourhood configuration to improve the KNN based imputation algorithms on microarray gene expression data

Due to technical problems in DNA microarray experiments, a large number of entries are found missing in microarray datasets. As a consequence, the effectiveness of the analysis algorithms deteriorates. Among different imputation techniques, the weighted average based methods always generate consistent results, are algorithmically simple and very popular but they also suffer from some drawbacks. These deficiencies have been pointed out in this work, and a new framework has been suggested to overcome those. The proposed framework is embedded in the K-nearest neighbour imputation method (KNNimpute), as well as its different versions. It is based on a hybrid distance and gene transformation procedure which utilises simultaneously the advantages of Euclidean distance, mean squared residue score, and Pearson correlation coefficient to select the best possible neighbours, using pattern-based similarity. The framework is tested on well-known microarray datasets. From the experimental results, the superiority of the proposed work has been found.

Dual Reconstructive Autoencoder for Crowd Localization and Estimation in Density and FIDT Maps

This paper proposes crowd estimation technology to help authorities make the right decisions in times of crisis. Specifically, deep learning models have faced these challenges, achieving excellent results. In particular, the trend of using single-column Fully Convolutional Networks (FCNs) has increased in recent years. A typical architecture that meets these characteristics is the autoencoder. However, this model presents an intrinsic difficulty: the search for the optimal dimensionality of the latent space. In order to alleviate such difficulty, we propose a dual architecture consisting of two cascaded autoencoders. The first autoencoder is responsible for carrying out the masked reconstruction of the original images, whereas the second obtains crowd maps from the outputs of the first one. In this way, our architecture improves the location of people and crowds in Focal Inverse Distance Transform (FIDT) maps, resulting in more reliable count estimates concerning a single autoencoder architecture.

Prediction method of dust pollutant diffusion range in building demolition based on Euclidean distance transformation

Prediction method of dust pollutant diffusion range in building demolition based on Euclidean distance transformation

Crowd Counting and Localization Beyond Density Map

Crowd analysis in general and counting in congested scenes, in particular, is an effective and vibrant research domain in computer vision due to its numerous applications. Understanding the risk analysis and safety aspects of crowd dynamics at various vital occasions related to sports, cultural and religious activities, specifically, Hajj and Umrah is essential. Thousands of people gathered in a small area to carry out their rites. Localizing and counting the annotated head points is quite challenging due to occlusion and large-scale variation in the congested environment. To deal with these problems, a small and effective solution is to generate the density maps. However, the significant flaws of the density map have a blurry Gaussian blob which is less effective for counting and localizing head annotations in the congested scene. To overcome these issues, we propose Congested Scene Crowd Counting and Localization Network (CSCCL-Net) with a Focal inverse Distance Transform (FIDT) map that can count and localize the people simultaneously in the highly congested scene. To evaluate the proposed model’s efficiency, extensive tests were performed on the ShanghaiTech part A, ShanghaiTech part B, and JHU-CROWD++ datasets. The proposed model outperforms existing state-of-the-art techniques regarding high accuracy and low Mean Absolute Error (MAE) and Mean Square Error (MSE) values.

Two-Stage Segmentation Framework Based on Distance Transformation.

With the rise of deep learning, using deep learning to segment lesions and assist in diagnosis has become an effective means to promote clinical medical analysis. However, the partial volume effect of organ tissues leads to unclear and blurred edges of ROI in medical images, making it challenging to achieve high-accuracy segmentation of lesions or organs. In this paper, we assume that the distance map obtained by performing distance transformation on the ROI edge can be used as a weight map to make the network pay more attention to the learning of the ROI edge region. To this end, we design a novel framework to flexibly embed the distance map into the two-stage network to improve left atrium MRI segmentation performance. Furthermore, a series of distance map generation methods are proposed and studied to reasonably explore how to express the weight of assisting network learning. We conduct thorough experiments to verify the effectiveness of the proposed segmentation framework, and experimental results demonstrate that our hypothesis is feasible.

Predicting the multi-label protein subcellular localization through multi-information fusion and MLSI dimensionality reduction based on MLFE classifier.

MotivationMulti-label (ML) protein subcellular localization (SCL) is an indispensable way to study protein function. It can locate a certain protein (such as the human transmembrane protein that promotes the invasion of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)) or expression product at a specific location in a cell, which can provide a reference for clinical treatment of diseases such as coronavirus disease 2019 (COVID-19).ResultsThe article proposes a novel method named ML-locMLFE. First of all, six feature extraction methods are adopted to obtain protein effective information. These methods include pseudo amino acid composition, encoding based on grouped weight, gene ontology, multi-scale continuous and discontinuous, residue probing transformation and evolutionary distance transformation. In the next part, we utilize the ML information latent semantic index method to avoid the interference of redundant information. In the end, ML learning with feature-induced labeling information enrichment is adopted to predict the ML protein SCL. The Gram-positive bacteria dataset is chosen as a training set, while the Gram-negative bacteria dataset, virus dataset, newPlant dataset and SARS-CoV-2 dataset as the test sets. The overall actual accuracy of the first four datasets are 99.23%, 93.82%, 93.24% and 96.72% by the leave-one-out cross validation. It is worth mentioning that the overall actual accuracy prediction result of our predictor on the SARS-CoV-2 dataset is 72.73%. The results indicate that the ML-locMLFE method has obvious advantages in predicting the SCL of ML protein, which provides new ideas for further research on the SCL of ML protein.Availability and implementationThe source codes and datasets are publicly available at https://github.com/QUST-AIBBDRC/ML-locMLFE/.Supplementary information Supplementary data are available at Bioinformatics online.

Saliency-guided deep learning network for automatic tumor bed volume delineation in post-operative breast irradiation

Efficient, reliable and reproducible target volume delineation is a key step in the effective planning of breast radiotherapy. However, post-operative breast target delineation is challenging as the contrast between the tumor bed volume (TBV) and normal breast tissue is relatively low in CT images. In this study, we propose to mimic the marker-guidance procedure in manual target delineation. We developed a saliency-based deep learning segmentation (SDL-Seg) algorithm for accurate TBV segmentation in post-operative breast irradiation. The SDL-Seg algorithm incorporates saliency information in the form of markers’ location cues into a U-Net model. The design forces the model to encode the location-related features, which underscores regions with high saliency levels and suppresses low saliency regions. The saliency maps were generated by identifying markers on CT images. Markers’ location were then converted to probability maps using a distance transformation coupled with a Gaussian filter. Subsequently, the CT images and the corresponding saliency maps formed a multi-channel input for the SDL-Seg network. Our in-house dataset was comprised of 145 prone CT images from 29 post-operative breast cancer patients, who received 5-fraction partial breast irradiation (PBI) regimen on GammaPod. The 29 patients were randomly split into training (19), validation (5) and test (5) sets. The performance of the proposed method was compared against basic U-Net. Our model achieved mean (standard deviation) of 76.4(±2.7) %, 6.76(±1.83) mm, and 1.9(±0.66) mm for Dice similarity coefficient, 95 percentile Hausdorff distance, and average symmetric surface distance respectively on the test set with computation time of below 11 seconds per one CT volume. SDL-Seg showed superior performance relative to basic U-Net for all the evaluation metrics while preserving low computation cost. The findings demonstrate that SDL-Seg is a promising approach for improving the efficiency and accuracy of the on-line treatment planning procedure of PBI, such as GammaPod based PBI.

IPromoter-ET: Identifying promoters and their strength by extremely randomized trees-based feature selection

Promoter is a region of DNA that determines the transcription of a particular gene. There are several σ factors in the RNA polymerase, which has the function of identifying the promoter and facilitating the binding of the RNA polymerase to the promoter. Owing to the importance of promoter in genome research, it is an urgent task to develop computational tool for effectively identifying promoters and their strength facing the avalanche of DNA sequences discovered in the post-genomic age. In this paper, we develop a model named iPromoter-ET using the k-mer nucleotide composition, binary encoding and dinucleotide property matrix-based distance transformation for features extraction, and extremely randomized trees (extra trees) for feature selection. Its 1st layer is used to identify whether a DNA sequence is of promoter or not, while its 2nd layer is to identify promoter samples as being strong or weak promoter. Support vector machine and the five cross-validation are used to perform identification and assess performance, respectively. The results indicate that our model remarkably outperforms the existing models in both the 1st and 2nd layers for accuracy and stability. We anticipate that our proposed model will become a very effective intelligent tool, or at the least, a complementary tool to the existing modes of identifying promoters and their strength. Moreover, the datasets and codes for iPromoter-ET are freely available at https://github.com/shengli0201/iPromoter-ET.

EUCLIDIAN DISTANCE TRANSFORMATION, MAIN AXIS ROTATION AND NOISY DILITATION SUPPORTED CROSS-SECTION CLASSIFICATION WITH CONVOLUTIONAL NEURAL NETWORKS

AbstractPolygon meshes and particularly triangulated meshes can be used to describe the shape of different types of geometry such as bicycles, bridges, or runways. In engineering, such polygon meshes can be supplied as finite element meshes, resulting from topology optimization or from laser scanning. Especially from topology optimization, frame-like polygon meshes with slender parts are typical and often have to be converted into a CAD (Computer-Aided Design) format, e.g., for further geometrical detailing or performing additional shape optimization. Especially for such frame-like geometries, CAD designs are constructed as beams with cross-sections and beam-lines, whereby the cross-section is extruded along the beam-lines or beam skeleton. One major task in the recognition of beams is the classification of the cross-section type such as I, U, or T, which is addressed in this article. Therefore, a dataset consisting of different cross-sections represented as binary images is created. Noisy dilatation, the distance transformation, and main axis rotation are applied to these images to increase the robustness and reduce the necessary amount of samples. The resulting images are applied to a convolutional neuronal network.

SUBDIVISION SURFACE MID-SURFACE RECONSTRUCTION OF TOPOLOGY OPTIMIZATION RESULTS AND THIN-WALLED SHAPES USING SURFACE SKELETONS

AbstractPolygon meshes and particularly triangulated meshes can be used to describe the shape of different types of geometry such as bicycles, bridges, or runways. In engineering, such polygon meshes can occur as finite element meshes, resulting from topology optimization or laser scanning. This article presents an automated parameterization of polygon meshes into a parametric representation using subdivision surfaces, especially in topology optimization. Therefore, we perform surface skeletonization on a volumetric grid supported by the Euclidian distance transformation and topology preserving and shape-preserving criterion. Based on that surface skeleton, an automated conversation into a Subdivision Surface Control grid is established. The final mid-surface-like parametrization is quite flexible and can be changed by variating the control gird or the local thickness.

An enhanced methodology for predicting protein-protein interactions between human and hepatitis C virus via ensemble learning algorithms

Hepatitis C virus (HCV) is responsible for a variety of human life-threatening diseases, which include liver cirrhosis, chronic hepatitis, fibrosis and hepatocellular carcinoma (HCC) . Computational study of protein-protein interactions between human and HCV could boost the findings of antiviral drugs in HCV therapy and might optimize the treatment procedures for HCV infections. In this analysis, we constructed a prediction model for protein-protein interactions between HCV and human by incorporating the features generated by pseudo amino acid compositions, which were then carried out at two levels: categories and features. In brief, extra-tree was initially used for feature selection while SVM was then used to build the classification model. After that, the most suitable models for each category and each feature were selected by comparing with the three ensemble learning algorithms, that is, Random Forest, Adaboost, and Xgboost. According to our results, profile-based features were more suitable for building predictive models among the four categories. AUC value of the model constructed by Xgboost algorithm on independent data set could reach 92.66%. Moreover, Distance-based Residue, Physicochemical Distance Transformation and Profile-based Physicochemical Distance Transformation performed much better among the 17 features. AUC value of the Adaboost classifier constructed by Profile-based Physicochemical Distance Transformation on the independent dataset achieved 93.74%. Taken together, we proposed a better model with improved prediction capacity for protein-protein interactions between human and HCV in this study, which could provide practical reference for further experimental investigation into HCV-related diseases in future. Communicated by Ramaswamy H. Sarma

Merging radar and rain gauge data by using spatial–temporal local weighted linear regression kriging for quantitative precipitation estimation

Merging radar and rain gauge data is an important means of obtaining precipitation products with high accuracy and high spatial–temporal resolution. However, due to the change of the raindrop size distribution itself and the uncertainty of the radar observational data, it is an indisputable fact that A and b in the Z–R (radar reflectivity factor (Z) to rainfall rate (R)) relationship are complex functions of space and time, which has a great influence on the accuracy of radar–rain gauge quantitative precipitation estimation (QPE). In this study, in an effort to further improve the accuracy of radar–rain gauge QPE, spatial–temporal local weighted linear regression (STLWLR) and its corresponding regression kriging (STLWLRK) are proposed for merging radar and rain gauge data for QPE, for which the maximum number of Z–R pairs in the spatial–temporal neighborhood N, the spatial–temporal distance transformation parameter μ, and the specified threshold C of |e/σ| are three key parameters. Specifically, cases that occurred on Hainan Island and the cross-validation mode were used to calibrate the three key parameters and to evaluate the two methods with the relative mean absolute error (RMAE) and bias as evaluation indicators. The results show that compared with the real-time-adjusted Z–R relationship (AZR) scheme and the two-step calibration technique of radar QPE (the AID scheme), STLWLR and STLWLRK with optimization parameters could further improve the accuracy of radar–rain gauge QPE. For 10-fold cross-validation of the cases occurring on Hainan Island from 1 May to 31 October 2018 and for 1-hr, 3-hr, 6-hr, 12-hr, and 24-hr time scales, the RMAE of STLWLRK decreased by 6.3%, 8.4%, 8.4%, 8.2%, and 7.9%, respectively, compared with that of the AID scheme. The results also show that increasing time scale could reduce the error of radar-rain gauge QPE for different methods. As STLWLR is based on solid geography, mathematics, and radar meteorology and has been tested by a large number of cases, STLWLR and its optimization parameters should have certain universality. Overall, compared with traditional methods, STLWLRK is a more robust, easier to implement, more accurate, and less computationally expensive method.

UMAP-DBP: An Improved DNA-Binding Proteins Prediction Method Based on Uniform Manifold Approximation and Projection.

DNA-binding proteins play a vital role in cellular processes. It is an extremely urgent to develop a high-throughput method for efficiently identifying DNA-binding proteins. According to the current research situation, some methods in machine learning and deep learning show excellent computational speed and accuracy, which are worthy of application. In this work, a novel predictor was proposed to predict DNA binding proteins called UMAP-DBP. Firstly, the feature extraction of primary protein sequence was realized based on physicochemical distance transformation, Profile-based auto-cross covariance and General series correlation pseudo amino acid composition. Secondly, uniform manifold approximation and projection (UMAP) and feature importance score methods were used for feature selection; there is a progressive relationship between them. Finally, the Adaboost operation engine with jackknife test were adopted for predicting DNA-binding proteins. For the jackknife test on the BP1075 and BP594, we obtained an overall accuracy of 82.97% and 82.14%, Cohen's kappa (CK) of 0.66 and 0.64, respectively. The results illustrate that a feasible method has been developed for predicting DNA-binding proteins by UMAP and Adaboost. This is the first study in which UMAP has been successfully applied to identify DNA-binding proteins. All the datasets and codes are accessible at https://github.com/Wang-Jinyue/UMAP-DBP .

On the role of distance transformations in Baddeley’s Delta Metric

Comparison and similarity measurement have been a key topic in computer vision for a long time. There is, indeed, an extensive list of algorithms and measures for image or subimage comparison. The superiority or inferiority of different measures is hard to scrutinize, especially considering the dimensionality of their parameter space and their many different configurations. In this work, we focus on the comparison of binary images, and study different variations of Baddeley’s Delta Metric, a popular metric for such images. We study the possible parameterizations of the metric, stressing the numerical and behavioural impact of different settings. Specifically, we consider the parameter settings proposed by the original author, as well as the substitution of distance transformations by regularized distance transformations, as recently presented by Brunet and Sills. We take a qualitative perspective on the effects of the settings, and also perform quantitative experiments on separability of datasets for boundary evaluation.