Boundary Detection Method Research Articles

Real-Time Mine Road Boundary Detection and Tracking for Autonomous Truck.

Road boundary detection is an important part of the perception of the autonomous driving. It is difficult to detect road boundaries of unstructured roads because there are no curbs. There are no clear boundaries on mine roads to distinguish areas within the road boundary line and areas outside the road boundary line. This paper proposes a real-time road boundary detection and tracking method by a 3D-LIDAR sensor. The road boundary points are extracted from the detected elevated point clouds above the ground point cloud according to the spatial distance characteristics and the angular features. Road tracking is to predict and update the boundary point information in real-time, in order to prevent false and missed detection. The experimental verification of mine road data shows the accuracy and robustness of the proposed algorithm.

Key Feature Extraction for Video Shot Boundary Detection using CNN

Now days as the progress of digital image technology, video files raise fast, there is a great demand for automatic video semantic study in many scenes, such as video semantic understanding, content-based analysis, video retrieval. Shot boundary detection is an elementary step for video analysis. However, recent methods are time consuming and perform badly in the gradual transition detection. In this paper we have projected a novel approach for video shot boundary detection using CNN which is based on feature extraction. We designed couple of steps to implement this method for automatic video shot boundary detection (VSBD). Primarily features are extracted using H, V&S parameters based on mean log difference along with implementation of histogram distribution function. This feature is given as an input to CNN algorithm which detects shots which is based on probability function. CNN is implemented using convolution and rectifier linear unit activation matrix which is followed after filter application and zero padding. After downsizing the matrix it is given as a input to fully connected layer which indicates shot boundaries comparing the proposed method with CNN method based on GPU the results are encouraging with substantially high values of precision Recall & F1 measures. CNN methods perform moderately better for animated videos while it excels for complex video which is observed in the results.

Flow and mixing characteristics of double-concentric jets pulsed at annular flow

The flow and mixing characteristics of double concentric jets pulsed at the annular flow were experimentally studied. The evolution process of the flow was observed and analyzed by means of the laser-light sheet assisted flow visualization method using the instantaneous streak pictures. The jet spread width was obtained using the binary boundary detection method based on the long-exposure pictures of the flow. The average velocities and turbulence intensities in both the centerline and radial directions were measured using the hot-wire anemometry. The mixing properties were determined using the tracer gas concentration detection technique. Three characteristic flow modes were identified in the domain of the central jet and the annular flow Reynolds numbers, namely the annular flow dominated mode, central jet dominated mode-low shear, and central jet dominated mode-high shear. In the annular flow dominated mode, the annular flow broke up the central jet in the centerline, where the axial velocity decayed quickly and the turbulence intensity increased to the maximum value. The larger jet spread in the radial direction enhanced the mixing characteristics in the recirculation region. In the central jet dominated mode-low shear, the central jet penetrated the recirculation region, resulting in significant entrainment of the annular flow. The axial velocity was modest and the turbulence intensity was insignificant; therefore, low mixing was observed in the recirculation region. In the central jet dominated mode-high shear, the recirculation bubble shrunk rapidly toward the jet exit, while the jet spread was controlled by the strong central jet entrainment. Although the annular flow interrupted the central jet and the turbulence intensity reached a peak after the combined jet, the enhancement of the mixing was quite seen to be lower than in the case of the annular flow dominated mode.

Static boundary detection of rail transit vehicles based on reference coordinate system

Aiming at the problems of low accuracy and large relative error of traditional rail transit vehicle static boundary detection method, a method of rail transit vehicle static boundary detection based on reference coordinate system is proposed. The parametric equations of measuring points are obtained by mathematical model. The model and boundary are measured by establishing reference orbit coordinate system based on laser tracking. By using self-calibration multi-sensor system calibration and DGPS/IMU integrated processing, the internal and external bits of each binocular image are calculated, and the relative relations of each element in the stereo image are constructed. Combining with the vehicle boundary, the vehicle boundary parameters are calculated to realise the static boundary detection of rail transit vehicles. Experimental results show that this method can effectively improve the accuracy of vehicle boundary detection and reduce the relative error of vehicle boundary detection.

An All-New Ultradeep Detection Method Based on Hybrid Dipole Antennas in Electromagnetic Logging While Drilling

We have developed an all-new ultradeep boundary detection method for reservoir mapping using borehole electromagnetic (EM) logging while drilling (LWD) measurements. The new method is based on hybrid dipole antennas, which combine the magnetic dipole (MD) transmitter with an electric dipole (ED) receiver. This takes advantage of the strong signal strength of the crosscoupling electric component and its slow attenuation property. Thus, the detection capability improves significantly. Moreover, the induced signal at the ED antenna usually enlarges with decreasing transmitter-to-receiver (T–R) spacing, which has balanced the tradeoff between the depth of detection (DoD) and the tool size. A practical geosteering tool configuration is then proposed by introducing a new open-loop half-circle antenna. Numerical results demonstrate that the new method is feasible for ultradeep boundary detection. By using the coaxial opened half-circle antenna and a coaxial MD transmitter, the DoD can be extended to tens of meters, up to 16 times larger than the T–R spacing under favorable conditions. The new method is a breakthrough in geosteering techniques and it will benefit the wellbore placement optimizations and reservoir evaluation.

Empirical wavelet transform based automated alcoholism detecting using EEG signal features

Electroencephalogram (EEG) signals are well used to characterize the brain states and actions. In this paper, a novel empirical wavelet transform (EWT) based machine learning framework is proposed for the classification of alcoholic and normal subjects using EEG signals. In the framework, the adaptive filtering is used to extract Time–Frequency-domain features from Hilbert–Huang Transform (HHT). The boundary detection method is used for segmenting the Fourier spectrum of the EEG signals to represent in scale-space. Hilbert–Huang Transform (HHT) examines time and frequency information in a single domain using instantaneous amplitude (IA) and instantaneous frequency (IF). The IA and IF are used to form intrinsic mode functions (IMF). The empirical wavelets transform (EWT) using Hilbert–Huang transforms (HHT) extract the statistical features such as mean, standard deviation, variance, skewness, kurtosis, Shannon entropy, and log entropy from each of the intrinsic mode functions (IMF). The extracted features are evaluated by t-test for finding the most significant features. The significant feature matrix is fed to various classification algorithms listed as least square-support vector machine (LS-SVM), support vector machine (SVM), Naïve Bayes (NB), and k-Nearest Neighbors (K-NN). The leave-one-out cross-validation (LOOCV) is used for training and testing of used models to minimize the chance of overfitting. The results suggest that the highest numbers of the positive samples are obtained using LS-SVM classifier with the polynomial kernel. The LS-SVM also achieved an average accuracy of 98.75%, the sensitivity 98.35%, specificity 99.16%, the precision 99.17%, F-measure 98.76%, and Matthews Correlation Coefficient (MCC) 97.50%.

Multi-time Window Automatic First-arrival Picking Method

The sensitivity of energy ratio method varies with the size of time window. We propose a multi-time window energy boundary detection method which improves the picking accuracy for data with a low-to-medium signal to noise ratio (SNR). The multi-time window algorithm effectively improves the system reliability of the energy ratio method. Through the distribution of characteristic values in different time windows, the picking results of low signal-to-noise ratio data can be effectively deleted. Then, a small-step fitting algorithm is applied to the remaining first-arrival characteristic values to obtain the final characteristic value evaluation. Based on the retained first-arrival characteristic values, the missing values were assigned by interpolation, then map the result on the original record and finally, first-arrival picking was completed by using a small time window. We test the performance of each module to prove the validity of the proposed methods. Finally, we design and develop auto-picking system, and the application shows that the accuracy of picking up can meet the needs of production.

Prosodic word boundary detection from Bengali continuous speech

Detection of word boundaries in continuous speech is a tedious process due to the absence of a definite pause or silence in the word boundary position. Thus, continuous speech recognition is a very challenging task. However, the prosodic word boundaries, unlike the written word boundaries, can be predicted using the prosodic parameters of continuous speech. This paper proposes a method for detecting such prosodic word boundaries from Bengali continuous speech. Bengali is a bound-stress language, where stress is observed on the first syllable of a prosodic word. Empirical Mode Decomposition is applied to the logarithm of fundamental frequency (F0) contour of continuous speech to detect prosodic word boundaries. 200 Bengali readout sentences, read by ten speakers, are analyzed for the present work. An overall prosodic boundary detection accuracy of 88.05% is achieved, whereas precision and recall values are 90.73% and 88.31%, respectively, with f-score as 89.5. A prosodic word dictionary comprising 5031 prosodic words has been developed by analyzing 1526 Bengali sentences with the proposed prosodic word boundary detection method.

A Proposed Method for Video's Shot Boundary Detection based on Multiple Features

A Proposed Method for Video's Shot Boundary Detection based on Multiple Features

Motion Boundary Detection Improved by Bio-Inspired Approach

Motion detection in video sequences is a problem that different techniques and algorithms have faced. Several computer vision applications require an efficient detection stage. A reliable motion detection system should work under a wide range of a variety of scenes and various acquisition conditions. In this paper, a new approach to detect moving objects boundaries in an image sequence is presented. The proposed approach is based on the frame differing of modified Haynes technique combined with a biologically inspired spiking neural network for boundaries detection. First, the detection of temporal changes consists in extracting the moving objects between two consecutive frames in a dynamic scene without any estimate of the optical flow, nor any priory on the stage as the background image. The main goal is to get a fast and effective detection. In order to obtain a motion boundaries information, the extracted moving objects are then presented to a spiking neural network based on the Hodgkin-Huxley neuron model whose responses can form a complete closed curve coinciding with the motion object boundaries. Simulation results in some experiments are satisfactory. The performance of the motion boundary detection system is compared against competing boundary detection method using the YouTube Motion Boundaries dataset (YMB) dataset. The comparison results obtained confirm the robustness of the proposed approach. The extracted boundaries objects results can be used in further applications.

An Efficient and Robust Temporal Video Segmentation

Temporal video segmentation is the primary step of content based video retrieval. The whole processes of video management are coming under the focus of content based video retrieval, which includes, video indexing, video retrieval, and video summarization etc. In this paper, we proposed a computationally efficient and discriminating shot boundary detection method, which uses a local feature descriptor named local Contrast and Ordering (LCO) for feature extraction. The results of the experiments, which are conducted on the video dataset TRECVid, analyzed and compared with some existing shot boundary detection methods. The proposed method has given a promising result, even in the cases of illumination changes, rotated images etc.

A Method of Scene Boundary Detection for Indexing Video Data Efficiently

In a cell animation, a background scene is presented with one cell. When animation scenes are changed, a relatively large scene change occurs because their backgrounds are also changed. Unlike a camera-based real movie, a cell animation is made in the way of manual drawing. For this reason, such animation does not use many colors. In order to apply the characteristics of animations as most as possible and detect a scene change of a cell animation effectively, this study proposes a new scene change detection technique with the stepwise use of the color and block-based histograms. The proposed algorithm receives continuous animation images as input, changes the RGB color space into the HSI color space, executes the difference operation of color values of the two images, and thereby primarily determines whether the neighboring images are scene change candidates. If they are judged to be scene change candidates, the color histogram for each sub-region is made, and then a weight value is applied to finally determine whether a scene change occurs. To compare the proposed scene change detection method with conventional scene change detection methods, the block-based scene change detection method and the histogram-based scene change detection method as conventional ones were implemented. In this study, to qualitatively evaluate the performance of the proposed method of detecting scene boundaries, the accuracy measure was defined. The two conventional algorithms are not complicated relatively and are widely used to extract a scene change. The experiment of this study reveals that the proposed method more accurately detects an animation scene change than other conventional methods. The proposed animation scene boundary detection method is expected to be applied usefully to the fields of digital video data indexing and retrieval, dynamic motion analysis, and other related areas.

Boundary particle resampling for surface reconstruction in liquid animation

In this paper, we present a novel adaptive particle resampling method tailored for surface reconstruction of level-sets defined by the boundary particles from a particle-based liquid simulation. The proposed approach is simple and easy to implement, and only requires the positions of the particles to identify and refine regions with small and thin fluid features accurately. The method comprises four main stages: boundary detection, feature classification, particle refinement, and surface reconstruction. For each simulation frame, firstly the free-surface particles are captured through a boundary detection method. Then, the boundary particles are classified and labeled according to the deformation and the stretching of the free-surface computed from the Principal Component Analysis (PCA) of the particle positions. The particles placed at feature regions are then refined according to their feature classification. Finally, we extract the free-surface of the zero level-set defined by the resampled boundary particles and its normals. In order to render the free-surface, we demonstrate how the traditional methods of surface fitting in Computer Graphics and Computational Physics literature can benefit from the proposed resampling method. Furthermore, the results shown in the paper attest the effectiveness and robustness of our method when compared to state-of-the-art adaptive particle resampling techniques.

IMAGE SEGMENTATION AND OBJECT SELECTION BASED ON MULTI-THRESHOLD PROCESSING

Introduction. Detection, isolation, selection and localization of variously shaped objects in images are essential in a variety of applications. Computer vision systems utilizing television and infrared cameras, synthetic aperture surveillance radars as well as laser and acoustic remote sensing systems are prominent examples. Such problems as object identification, tracking and matching as well as combining information from images available from different sources are essential. Objective. Design of image segmentation and object selection methods based on multi-threshold processing. Materials and methods. The segmentation methods are classified according to the objects they deal with, including (i) pixel-level threshold estimation and clustering methods, (ii) boundary detection methods, (iii) regional level, and (iv) other classifiers, including many non-parametric methods, such as machine learning, neural networks, fuzzy sets, etc. The keynote feature of the proposed approach is that the choice of the optimal threshold for the image segmentation among a variety of test methods is carried out using a posteriori information about the selection results. Results. The results of the proposed approach is compared against the results obtained using the well-known binary integration method. The comparison is carried out both using simulated objects with known shapes with additive synthesized noise as well as using observational remote sensing imagery. Conclusion. The article discusses the advantages and disadvantages of the proposed approach for the selection of objects in images, and provides recommendations for their use.

A novel shot boundary detection system using hybrid optimization technique

This paper proposes a novel shot boundary detection method which combines the Advantages of Particle Swarm Optimization (PSO) and Gravitational Search Algorithm (GSA) to optimize the weights of the Feed-Forward Neural Network (FNN). To increase the performance of the system, the output of the hybrid technique is again analyzed by forming a Continuity matrix (ϕ). Then an Outlier along with a Continuity matrix is used for extracting a possible set of transition frames. A set of thresholds δ1 and δ2 is selected for classifying abrupt and gradual transitions from the available set of possible transition frames. Experimental results using TRECVid 2001 depicts that PSOGSA outperforms GSA and PSO in-terms of training the feed forward neural network and generating a higher overall F1 score. The proposed system also gives better performance when compared with other latest techniques in-terms of F1 score.

Two Stage Shot Boundary Detection via Feature Fusion and Spatial-Temporal Convolutional Neural Networks

Shot boundary detection is essentially to detect the position of frames where the shot changes. It has been actively studied in video analysis and management for convenience, which becomes a key technique with the rapid proliferation of rich and diverse videos. With respect to the complex characteristics of different shots in varying length and content variation property, in this paper we present a two stage method for shot boundary detection (TSSBD) which distinguishes abrupt shot by fusing color histogram and deep features, and locate gradual shot changes with C3D-based deep analysis. Abrupt shot changes are detected firstly as it occurs between two frames, which divides the complete video into segments containing gradual transitions; Over these video segments, gradual shot change detection is implemented using 3D-convolutional neural network, which classifies clips into specific gradual shot change types; Finally, an effective merging strategy is proposed to locate positions of gradual shot transitions. The experimental analysis illustrates that the proposed progressive method is capable of detecting both abrupt shot transitions and gradual shot transitions accurately.

ERP/MMR Algorithm for Classifying Topic‐Relevant and Topic‐Irrelevant Visual Shots of Documentary Videos

We propose and evaluate a video summarization method based on a topic relevance model, a maximal marginal relevance (MMR), and discriminant analysis to generate a semantically meaningful video skim. The topic relevance model uses event‐related potential (ERP) components to describe the process of topic relevance judgment. More specifically, the topic relevance model indicates that N400 and P600, which have been successfully applied to the mismatch process of a stimulus and the discourse‐internal reorganization and integration process of a stimulus, respectively, are used for the topic mismatch process of a topic‐irrelevant video shot and the topic formation process of a topic‐relevant video shot. To evaluate our proposed ERP/MMR‐based method, we compared the video skims generated by the ERP/MMR‐based, ERP‐based, and shot boundary detection (SBD) methods with ground truth skims. The results showed that at a significance level of 0.05, the ROUGE‐1 scores of the ERP/MMR method are statistically higher than those of the SBD method, and the diversity scores of the ERP/MMR method are statistically higher than those of the ERP method. This study suggested that the proposed method may be applied to the construction of a video skim without operational intervention, such as the insertion of a black screen between video shots.

Automatic first-arrival picking method based on an image connectivity algorithm and multiple time windows

The use of a computer to automatically pick the first-arrival of a seismic signal is an operation that involves picking and screening the first arrival of the wave according to the criteria established in the manual picking process. To increase the picking accuracy for data with low-to-moderate signal-to-noise ratio (SNR), we propose a new single-trace boundary detection algorithm. This algorithm includes three steps: (1) calculate the first-arrival characteristic values through multi time windows; (2) take the times corresponding to the maximum characteristic values given by different time windows as intermediate results; (3) compare the intermediate results: if the difference is too large, it is marked the time is abnormal, otherwise the average time of the intermediate results is taken as the first-arrival time. Using this energy boundary detection method, the characteristic values obtained are bi-directionally expanded to allow the use of the trace connectivity algorithm which is improved from the region growing method. Determining the connectivity between the first-arrival characteristic values is a way to simulate how the human eye discriminates true first arrivals. This method significantly improves the elimination of false or abnormal first-arrivals. Next, a small-step fitting algorithm is applied to the remaining first-arrival characteristic values to complete the calculation of the final characteristic values. Based on the retained first-arrival characteristic values, the missing values are assigned by interpolation. The characteristic values are mapped on the original record and finally the first-arrival picking is completed using a small time window. Theoretical results as well as the results obtained from real data demonstrate that the proposed automatic first-arrival picking method effectively improves the accuracy of the first-arrival picking. Finally, the new picking algorithm is presented more efficient than the energy ratio method, as well as cross-correlation method.

A Study on Boundary Detection Method for Separation of Coastal Seahorse Community Obtained by Drones

A Study on Boundary Detection Method for Separation of Coastal Seahorse Community Obtained by Drones

Video Searching and Fingerprint Detection by Using the Image Query and PlaceNet-Based Shot Boundary Detection Method

This work presents a novel shot boundary detection (SBD) method based on the Place-centric deep network (PlaceNet), with the aim of using video shots and image queries for video searching (VS) and fingerprint detection. The SBD method has three stages. In the first stage, we employed Local Binary Pattern-Singular Value Decomposition (LBP-SVD) features for candidate shot boundaries selection. In the second stage, we used the PlaceNet to select the shot boundary by semantic labels. In the third stage, we used the Scale-Invariant Feature Transform (SIFT) descriptor to eliminate falsely detected boundaries. The experimental results show that our SBD method is effective on a series of SBD datasets. In addition, video searching experiments are conducted by using one query image instead of video sequences. The results under several image transitions by using shot fingerprints have shown good precision.