scholarly journals Detection and localization of an underwater docking station in acoustic images using machine learning and generalized fuzzy hough transform

2020 ◽  
pp. 23-31
Author(s):  
Divas Karimanzira ◽  
Helge Renkewitz
Keyword(s):  
Hough Transform ◽  
Template Matching ◽  
Autonomous Underwater Vehicle ◽  
Computation Time ◽  
Speckle Noise ◽  
Detection Methods ◽  
Acoustic Image ◽  
Window Approach ◽  
And Shifts ◽  
Detection And Localization

AbstractLong underwater operations with autonomous battery charging and data transmission require an Autonomous Underwater Vehicle (AUV) with docking capability, which in turn presume the detection and localization of the docking station. Object detection and localization in sonar images is a very difficult task due to acoustic image problems such as, non-homogeneous resolution, non-uniform intensity, speckle noise, acoustic shadowing, acoustic reverberation and multipath problems. As for detection methods which are invariant to rotations, scale and shifts, the Generalized Fuzzy Hough Transform (GFHT) has proven to be a very powerful tool for arbitrary template detection in a noisy, blurred or even a distorted image, but it is associated with a practical drawback in computation time due to sliding window approach, especially if rotation and scaling invariance is taken into account. In this paper we use the fact that the docking station is made out of aluminum profiles which can easily be isolated using segmentation and classified by a Support Vector Machine (SVM) to enable selective search for the GFHT. After identification of the profile locations, GFHT is applied selectively at these locations for template matching producing the heading and position of the docking station. Further, this paper describes in detail the experiments that validate the methodology.

scholarly journals Effective Opinion Spam Detection: A Study on Review Metadata Versus Content

2020 ◽  
Vol 5 (2) ◽  
pp. 76-110
Author(s):  
Ajay Rastogi ◽  
Monica Mehrotra ◽  
Syed Shafat Ali
Keyword(s):  
Computation Time ◽  
Detection Methods ◽  
Spam Detection ◽  
Time Analysis ◽  
Hybrid Features ◽  
Versus Content ◽  
Behavioral Features ◽  
Opinion Spam ◽  
Real World Datasets ◽  
Practical Implications

AbstractPurposeThis paper aims to analyze the effectiveness of two major types of features—metadata-based (behavioral) and content-based (textual)—in opinion spam detection.Design/methodology/approachBased on spam-detection perspectives, our approach works in three settings: review-centric (spam detection), reviewer-centric (spammer detection) and product-centric (spam-targeted product detection). Besides this, to negate any kind of classifier-bias, we employ four classifiers to get a better and unbiased reflection of the obtained results. In addition, we have proposed a new set of features which are compared against some well-known related works. The experiments performed on two real-world datasets show the effectiveness of different features in opinion spam detection.FindingsOur findings indicate that behavioral features are more efficient as well as effective than the textual to detect opinion spam across all three settings. In addition, models trained on hybrid features produce results quite similar to those trained on behavioral features than on the textual, further establishing the superiority of behavioral features as dominating indicators of opinion spam. The features used in this work provide improvement over existing features utilized in other related works. Furthermore, the computation time analysis for feature extraction phase shows the better cost efficiency of behavioral features over the textual.Research limitationsThe analyses conducted in this paper are solely limited to two well-known datasets, viz., YelpZip and YelpNYC of Yelp.com.Practical implicationsThe results obtained in this paper can be used to improve the detection of opinion spam, wherein the researchers may work on improving and developing feature engineering and selection techniques focused more on metadata information.Originality/valueTo the best of our knowledge, this study is the first of its kind which considers three perspectives (review, reviewer and product-centric) and four classifiers to analyze the effectiveness of opinion spam detection using two major types of features. This study also introduces some novel features, which help to improve the performance of opinion spam detection methods.

scholarly journals ACF Based Region Proposal Extraction for YOLOv3 Network Towards High-Performance Cyclist Detection in High Resolution Images

Sensors ◽  
2019 ◽  
Vol 19 (12) ◽  
pp. 2671 ◽  
Author(s):  
Chunsheng Liu ◽  
Yu Guo ◽  
Shuang Li ◽  
Faliang Chang
Keyword(s):  
High Resolution ◽  
High Precision ◽  
Extraction Method ◽  
High Performance ◽  
Detection Methods ◽  
Average Precision ◽  
Deep Network ◽  
High Resolution Images ◽  
Bounding Boxes ◽  
Detection And Localization

You Only Look Once (YOLO) deep network can detect objects quickly with high precision and has been successfully applied in many detection problems. The main shortcoming of YOLO network is that YOLO network usually cannot achieve high precision when dealing with small-size object detection in high resolution images. To overcome this problem, we propose an effective region proposal extraction method for YOLO network to constitute an entire detection structure named ACF-PR-YOLO, and take the cyclist detection problem to show our methods. Instead of directly using the generated region proposals for classification or regression like most region proposal methods do, we generate large-size potential regions containing objects for the following deep network. The proposed ACF-PR-YOLO structure includes three main parts. Firstly, a region proposal extraction method based on aggregated channel feature (ACF) is proposed, called ACF based region proposal (ACF-PR) method. In ACF-PR, ACF is firstly utilized to fast extract candidates and then a bounding boxes merging and extending method is designed to merge the bounding boxes into correct region proposals for the following YOLO net. Secondly, we design suitable YOLO net for fine detection in the region proposals generated by ACF-PR. Lastly, we design a post-processing step, in which the results of YOLO net are mapped into the original image outputting the detection and localization results. Experiments performed on the Tsinghua-Daimler Cyclist Benchmark with high resolution images and complex scenes show that the proposed method outperforms the other tested representative detection methods in average precision, and that it outperforms YOLOv3 by 13.69 % average precision and outperforms SSD by 25.27 % average precision.

Intrusion Detection of Cyber-Physical Attacks in Manufacturing Systems: A Review

2019 ◽  
Author(s):  
Mingtao Wu ◽  
Young B. Moon
Keyword(s):  
Intrusion Detection ◽  
Manufacturing Systems ◽  
High Volume ◽  
Cyber Attacks ◽  
Intrusion Detection Systems ◽  
Detection Methods ◽  
False Alarms ◽  
Alert Correlation ◽  
Acoustic Image ◽  
Detection Systems

Abstract Cyber-physical manufacturing system is the vision of future manufacturing systems where physical components are fully integrated through various networks and the Internet. The integration enables the access to computation resources that can improve efficiency, sustainability and cost-effectiveness. However, its openness and connectivity also enlarge the attack surface for cyber-attacks and cyber-physical attacks. A critical challenge in defending those attacks is that current intrusion detection methods cannot timely detect cyber-physical attacks. Studies showed that the physical detection provides a higher accuracy and a shorter respond time compared to network-based or host-based intrusion detection systems. Moreover, alert correlation and management methods help reducing the number of alerts and identifying the root cause of the attack. In this paper, the intrusion detection research relevant to cyber-physical manufacturing security is reviewed. The physical detection methods — using side-channel data, including acoustic, image, acceleration, and power consumption data to disclose attacks during the manufacturing process — are analyzed. Finally, the alert correlation methods — that manage the high volume of alerts generated from intrusion detection systems via logical relationships to reduce the data redundancy and false alarms — are reviewed. The study show that the cyber-physical attacks are existing and rising concerns in industry. Also, the increasing efforts in cyber-physical intrusion detection and correlation research can be utilized to secure the future manufacturing systems.

Sparse Extended Information Filter for AUV SLAM: Insights into the Optimal Sparse Time

2013 ◽  
Vol 427-429 ◽  
pp. 1670-1673 ◽  
Author(s):  
Hao Zhang ◽  
Bo He ◽  
Ning Luan
Keyword(s):  
Autonomous Underwater Vehicle ◽  
Computation Time ◽  
Underwater Vehicle ◽  
Information Loss ◽  
Sea Trial ◽  
Time Consumption ◽  
Localization And Mapping ◽  
Information Filter ◽  
Slam Algorithm ◽  
And Storage

Sparse extended information filter-based simultaneous localization and mapping (SEIF-based SLAM) algorithm can reflect significant advantages in terms of computation time and storage memories. However, SEIF-SLAM is easily prone to overconfidence due to sparsification strategy. In this paper we will consider the time consumption and information loss of sparse operation, and get the optimal sparse time. In order to verify the feasibility of sparsification, a sea trial for autonomous underwater vehicle (AUV) C-Ranger was conducted in Tuandao Bay. The experimental results will show the improved algorithm is much more effective and accurate comparedwithothermethods.

Non-destructive detection methods for discontinuities of interconnection structures

2016 ◽  
Vol 33 (1) ◽  
pp. 23-35 ◽  
Author(s):  
Tomas Blecha
Keyword(s):  
Transmission Lines ◽  
Printed Circuit Boards ◽  
Characteristic Impedance ◽  
Diagnostic Methods ◽  
Detection Methods ◽  
Circuit Boards ◽  
Content Type ◽  
Multiple Line ◽  
Detection And Localization ◽  
Non Destructive

Purpose – The purpose of this paper is to demonstrate the non-destructive methods for detection and localization of interconnection structure discontinuities based on the signal analysis in the frequency and time domain. Design/methodology/approach – The paper deals with the discontinuity characterization of interconnection structures created on substrates used for electronics, and methods for their detection and localization, based on the frequency analysis of transmitted signals. Used analyses are based on the theoretical approach for the solution of discontinuity electrical parameters and are the base for diagnostic methods of discontinuity identification. Findings – The measurement results of reflection parameters, frequency spectrums of transmitted signals and characteristic impedance values are presented on test samples containing multiple line cracks and their width reduction. Practical implications – Obtained results can be used practically, not only for the detection of transmission lines discontinuities on printed circuit boards but also in other applications, such as the quality assessment of bonded joints. Originality/value – Developed methods allow the quick identification and localization of particular discontinuities without the destruction of tested devices.

Edited film alignment via selective Hough transform and accurate template matching

2017 ◽  
Author(s):  
Xiaomeng Wu ◽  
Takahito Kawanishi ◽  
Minoru Mori ◽  
Kaoru Hiramatsu ◽  
Kunio Kashino
Keyword(s):  
Hough Transform ◽  
Template Matching

scholarly journals A Novel Segmentation Techniques for Red Blood Cells using Clustering Algorithms

2020 ◽  
Vol 9 (3) ◽  
pp. 2452-2460
Keyword(s):  
Cancer Cell ◽  
Blood Cells ◽  
Clustering Algorithms ◽  
Computation Time ◽  
Roc Curves ◽  
Extraction Process ◽  
Detection Methods ◽  
Novel Method ◽  
Serious Disease ◽  
Systems Effectiveness

RBC called Erythrocytes is one of the important element in blood composition which is main responsible in all living cells for its gaseous exchanges with the environment externally. In general, at the physiological maintained conditions, RBC in view provides circular in the front and also looks bi-concave at side. One of serious disease with reference to blood cells is Cancer where the healthy RBC are affected. This reduces the body's immunity factors. To identify the cancer cell various methods are employed but it does not provide the proper detection of blood cells. In this method, proper identification of the cancer cells from the unaffected RBCs was identified in which are presented in blood samples using various imaging tools and also with the techniques. The proposed novel method called Online Region Based Segmentation (ORBS) method is done which is used to discover the areas of the boundary of the unaffected corpuscles. By using properties of region, a suitable metric is formulated to determine the shape which is abnormal in the blood cells. Overall accuracy of 96.9% is obtained using proposed ORBS methods and deep learning classification (DLC) method is accurate as 97.1% that helps to diagnose cancer cell using the feature extraction process which is done automatically. The computation time was found to be less when related to the other existing method which is 22 seconds. Closeness of Proposed method in relative to True Positive values at ROC curves indicates the performance which is higher than other methods. Experimental results prove proposed systems effectiveness when compared by means of other detection methods.

scholarly journals Evaluation of Unsupervised Anomaly Detection Methods in Sentiment Mining

2019 ◽  
Vol 8 (9) ◽  
pp. 1080-1085
Keyword(s):  
Anomaly Detection ◽  
Computation Time ◽  
Vital Role ◽  
Detection Methods ◽  
Research Focus ◽  
Comparative Performance ◽  
Detection Techniques ◽  
Detection Algorithms ◽  
Sentiment Mining ◽  
Unsupervised Anomaly Detection

Anomaly detection has vital role in data preprocessing and also in the mining of outstanding points for marketing, network sensors, fraud detection, intrusion detection, stock market analysis. Recent studies have been found to concentrate more on outlier detection for real time datasets. Anomaly detection study is at present focuses on the expansion of innovative machine learning methods and on enhancing the computation time. Sentiment mining is the process to discover how people feel about a particular topic. Though many anomaly detection techniques have been proposed, it is also notable that the research focus lacks a comparative performance evaluation in sentiment mining datasets. In this study, three popular unsupervised anomaly detection algorithms such as density based, statistical based and cluster based anomaly detection methods are evaluated on movie review sentiment mining dataset. This paper will set a base for anomaly detection methods in sentiment mining research. The results show that density based (LOF) anomaly detection method suits best for the movie review sentiment dataset.

scholarly journals An Efficient and Fast Partial Template Matching Technique – Enhancement in Normalized Cross Correlation

2019 ◽  
Vol 9 (1) ◽  
pp. 7232-7237
Keyword(s):  
Image Processing ◽  
Execution Time ◽  
Template Matching ◽  
Absolute Difference ◽  
Processing Efficiency ◽  
Memory Space ◽  
Matching Technique ◽  
Speed Up ◽  
Novel Method ◽  
Window Approach

Template matching forms the basis of many image processing algorithms and hence the computer vision algorithms. There are many existing template matching algorithms like Sum of Absolute Difference (SAD), Normalized SAD (NSAD), Correlation methods (CORR), Normalized CORR(NCORR), Sum of Squared Difference (SSD), and Normalized SSD(NSSD). In general, as image requires more memory space for storage and much time for processing. The above said methods involves much computation. In any processing, efficiency constraints include many factors, especially accuracy of the results and speed of processing. An approach to reduce the execution time is always most appreciated. As a result of this, a novel method of partial NCC (PNCC) template matching technique is proposed in this paper. A block window approach is used to reduce the number of operations and hence to speed up the processing. A comparative study between existing NCC algorithm and the proposed partial NCC, PNCC algorithm is done. It is experimented and results proves that the execution time is reduced by 8 - 47 times approximately based on the various template images for different main images in PNCC. The accuracy of the result obtained is 100%. This proposed algorithm works for various types of images. The experiment is repeated for various sizes of templates and different sizes of main image. Further improvement in the speed of execution can be achieved by implementation of the proposed algorithm using parallel processors. It may find its importance in the real time image processing

Sign in / Sign up

Export Citation Format

Share Document