scholarly journals Evaluating effects of focal length and viewing angle in a comparison of recent face landmark and alignment methods

2021 ◽  
Vol 2021 (1) ◽  
Author(s):  
Xiang Li ◽  
Jianzheng Liu ◽  
Jessica Baron ◽  
Khoa Luu ◽  
Eric Patterson
Keyword(s):  
Focal Length ◽  
Ground Truth ◽  
Detection Methods ◽  
Viewing Angle ◽  
Deep Convolutional Neural Networks ◽  
Landmark Detection ◽  
High Performing ◽  
Detection Techniques ◽  
And Performance ◽  
Lens Focal Length

AbstractRecent attention to facial alignment and landmark detection methods, particularly with application of deep convolutional neural networks, have yielded notable improvements. Neither these neural-network nor more traditional methods, though, have been tested directly regarding performance differences due to camera-lens focal length nor camera viewing angle of subjects systematically across the viewing hemisphere. This work uses photo-realistic, synthesized facial images with varying parameters and corresponding ground-truth landmarks to enable comparison of alignment and landmark detection techniques relative to general performance, performance across focal length, and performance across viewing angle. Recently published high-performing methods along with traditional techniques are compared in regards to these aspects.

scholarly journals Identification of Tomato Disease Types and Detection of Infected Areas Based on Deep Convolutional Neural Networks and Object Detection Techniques

2019 ◽  
Vol 2019 ◽  
pp. 1-15 ◽  
Author(s):  
Qimei Wang ◽  
Feng Qi ◽  
Minghe Sun ◽  
Jianhua Qu ◽  
Jie Xue
Keyword(s):  
Neural Networks ◽  
Object Detection ◽  
Convolutional Neural Networks ◽  
Detection Methods ◽  
Disease Detection ◽  
Deep Convolutional Neural Networks ◽  
Detection Techniques ◽  
Tomato Diseases ◽  
Validation Set ◽  
Tomato Disease

This study develops tomato disease detection methods based on deep convolutional neural networks and object detection models. Two different models, Faster R-CNN and Mask R-CNN, are used in these methods, where Faster R-CNN is used to identify the types of tomato diseases and Mask R-CNN is used to detect and segment the locations and shapes of the infected areas. To select the model that best fits the tomato disease detection task, four different deep convolutional neural networks are combined with the two object detection models. Data are collected from the Internet and the dataset is divided into a training set, a validation set, and a test set used in the experiments. The experimental results show that the proposed models can accurately and quickly identify the eleven tomato disease types and segment the locations and shapes of the infected areas.

Characterizations of Diagnostic Properties and Detection Techniques of Fentanyl and Related Synthetic Opioids

2018 ◽  
Author(s):  
M. Shafiqur Rahman ◽  
Uttam K. Chakravarty
Keyword(s):  
Chemical Properties ◽  
Detection Methods ◽  
Drug Detection ◽  
Detection Techniques ◽  
Chemical Reagents ◽  
Synthetic Opioids ◽  
Detection Technology ◽  
Test Kit ◽  
Microfluidic Analysis ◽  
And Performance

Fentanyl, a synthetic opioid, is an extremely fast-acting synthetic narcotic analgesic having a high potency of approximately 100 to 200 times that of Morphine. As the synthetic opioid crisis continues to sweep across the world, detection technologies are required to be enhanced to detect, categorize, and identify synthetic opioids effectively. To detect fentanyl and its analogues, immunoassay screens are commercially available for urine specimens. Simple colorimetric assays, i.e., spot tests with paper strips, offer speed, simplicity of operation, portability, and affordability. The microfluidic behavior of the paper strips along with the properties of chemical reagents play significant role in drug detection methods. Therefore, the objectives of this study are to characterize the chemical properties of fentanyl and its analogues and to conduct microfluidic analysis for design optimization and performance evaluation of the fentanyl test strips. A computational fluid dynamics model is developed to evaluate the microfluidic properties. Analytical study and Experiments with test-kit samples are also conducted to validate the results of the numerical simulation. Finally, the performance parameters based on microfluidic analysis were reported showing the room for improvements in the detection technology of fentanyl and related synthetic opioids.

scholarly journals Indirectly Supervised Anomaly Detection of Clinically Meaningful Health Events from Smart Home Data

2021 ◽  
Vol 12 (2) ◽  
pp. 1-18
Author(s):  
Jessamyn Dahmen ◽  
Diane J. Cook
Keyword(s):  
Anomaly Detection ◽  
Time Series Data ◽  
Bayesian Optimization ◽  
Detection Methods ◽  
Series Data ◽  
Detection Techniques ◽  
Health Events ◽  
Warm Start ◽  
Health Related ◽  
Unsupervised Algorithms

Anomaly detection techniques can extract a wealth of information about unusual events. Unfortunately, these methods yield an abundance of findings that are not of interest, obscuring relevant anomalies. In this work, we improve upon traditional anomaly detection methods by introducing Isudra, an Indirectly Supervised Detector of Relevant Anomalies from time series data. Isudra employs Bayesian optimization to select time scales, features, base detector algorithms, and algorithm hyperparameters that increase true positive and decrease false positive detection. This optimization is driven by a small amount of example anomalies, driving an indirectly supervised approach to anomaly detection. Additionally, we enhance the approach by introducing a warm-start method that reduces optimization time between similar problems. We validate the feasibility of Isudra to detect clinically relevant behavior anomalies from over 2M sensor readings collected in five smart homes, reflecting 26 health events. Results indicate that indirectly supervised anomaly detection outperforms both supervised and unsupervised algorithms at detecting instances of health-related anomalies such as falls, nocturia, depression, and weakness.

A review on cephalometric landmark detection techniques

2021 ◽  
Vol 66 ◽  
pp. 102486
Author(s):  
Mamta Juneja ◽  
Poojita Garg ◽  
Ravinder Kaur ◽  
Palak Manocha ◽  
Prateek ◽  
...  
Keyword(s):  
Landmark Detection ◽  
Detection Techniques

scholarly journals Standalone Behaviour-Based Attack Detection Techniques for Distributed Software Systems via Blockchain

2021 ◽  
Vol 11 (12) ◽  
pp. 5685
Author(s):  
Hosam Aljihani ◽  
Fathy Eassa ◽  
Khalid Almarhabi ◽  
Abdullah Algarni ◽  
Abdulaziz Attaallah
Keyword(s):  
Attack Detection ◽  
Detection Methods ◽  
Software Systems ◽  
Distributed Software ◽  
Detection Techniques ◽  
Detection Systems ◽  
Novel Approach ◽  
Validation Experiment ◽  
Critical Issues ◽  
Concrete Solution

With the rapid increase of cyberattacks that presently affect distributed software systems, cyberattacks and their consequences have become critical issues and have attracted the interest of research communities and companies to address them. Therefore, developing and improving attack detection techniques are prominent methods to defend against cyberattacks. One of the promising attack detection methods is behaviour-based attack detection methods. Practically, attack detection techniques are widely applied in distributed software systems that utilise network environments. However, there are some other challenges facing attack detection techniques, such as the immutability and reliability of the detection systems. These challenges can be overcome with promising technologies such as blockchain. Blockchain offers a concrete solution for ensuring data integrity against unauthorised modification. Hence, it improves the immutability for detection systems’ data and thus the reliability for the target systems. In this paper, we propose a design for standalone behaviour-based attack detection techniques that utilise blockchain’s functionalities to overcome the above-mentioned challenges. Additionally, we provide a validation experiment to prove our proposal in term of achieving its objectives. We argue that our proposal introduces a novel approach to develop and improve behaviour-based attack detection techniques to become more reliable for distributed software systems.

State of Research on Negative Pressure Techniques Applied to Leak Detection in Liquid Pipelines

2004 ◽  
Author(s):  
Wei Liang ◽  
Lai-bin Zhang ◽  
Zhao-hui Wang
Keyword(s):  
Pressure Wave ◽  
Analytical Techniques ◽  
Leak Detection ◽  
Point Of View ◽  
Detection Methods ◽  
Propagation Mechanism ◽  
Detection Techniques ◽  
Wave Detection ◽  
State Of Research ◽  
Signal Processing Methods

In China, the rarefaction-pressure wave techniques are widely used to diagnose the leakage fault for liquid pipelines. Many leaking propagating assumptions, such as stable single-phased flow hypothesis and none rarefaction wave front hypothesis, are often uncertain in the process of leak detection, which can easily result in some errors. Thus the rarefaction-pressure wave techniques should be integrated with other analytical techniques to compute a more accurate leak location. Additionally, the development trends of rarefaction-pressure wave techniques lie in three aspects. First, rarefaction-pressure wave detection techniques will be integrated with other compatible detection techniques and modern signal processing methods to solve the complex problems encountered in leak detection. Second, studies of rarefaction-pressure wave techniques have advanced to a new stage. The deductions on propagation mechanism of rarefaction-pressure wave have been successfully applied to determine leaks qualitatively. Third, analysis on rarefaction-pressure wave detection techniques will be made from a quantitative point of view. The quantitative data have been used to deduce leak amounts and location. The purpose of this paper is to present the recent achievements in the study of improved rarefaction-pressure wave detection techniques. The rarefaction-pressure wave detection methods, effects of incomplete information conditions, the improvements of rarefaction-pressure wave detection techniques with modified factors and propagation mechanisms are comprehensively investigated. The disfigurements of rarefaction-pressure wave are analyzed. The corresponding methods for resolving such problems as ill diagnostic information and weak amplitude values are put forward. Several methods for stronger small leakage detection ability, higher leakage positioning precision, lower false alarm rates are proposed. The application of rarefaction-pressure wave detection techniques to safety protection of liquid pipelines is also introduced. Finally, the prospect of rarefaction-pressure wave detection techniques is predicted.

A virtual test environment for validating spacecraft optical navigation

2013 ◽  
Vol 117 (1197) ◽  
pp. 1075-1101 ◽  
Author(s):  
S. M. Parkes ◽  
I. Martin ◽  
M. N. Dunstan ◽  
N. Rowell ◽  
O. Dubois-Matra ◽  
...  
Keyword(s):  
Small Body ◽  
European Space Agency ◽  
Ground Truth ◽  
Camera Motion ◽  
Test Environment ◽  
Space Applications ◽  
Vision Guidance ◽  
Space Agency ◽  
Wide Range ◽  
And Performance

Abstract The use of machine vision to guide robotic spacecraft is being considered for a wide range of missions, such as planetary approach and landing, asteroid and small body sampling operations and in-orbit rendezvous and docking. Numerical simulation plays an essential role in the development and testing of such systems, which in the context of vision-guidance means that realistic sequences of navigation images are required, together with knowledge of the ground-truth camera motion. Computer generated imagery (CGI) offers a variety of benefits over real images, such as availability, cost, flexibility and knowledge of the ground truth camera motion to high precision. However, standard CGI methods developed for terrestrial applications lack the realism, fidelity and performance required for engineering simulations. In this paper, we present the results of our ongoing work to develop a suitable CGI-based test environment for spacecraft vision guidance systems. We focus on the various issues involved with image simulation, including the selection of standard CGI techniques and the adaptations required for use in space applications. We also describe our approach to integration with high-fidelity end-to-end mission simulators, and summarise a variety of European Space Agency research and development projects that used our test environment.

Deepfake videos: synthesis and detection techniques –a survey

2021 ◽  
pp. 1-21
Author(s):  
Shahela Saif ◽  
Samabia Tehseen
Keyword(s):  
Social Systems ◽  
Generative Models ◽  
Detection Methods ◽  
Future Research ◽  
Fake News ◽  
Research Directions ◽  
Detection Techniques ◽  
Future Research Directions ◽  
Video Creation ◽  
The One

Deep learning has been used in computer vision to accomplish many tasks that were previously considered too complex or resource-intensive to be feasible. One remarkable application is the creation of deepfakes. Deepfake images change or manipulate a person’s face to give a different expression or identity by using generative models. Deepfakes applied to videos can change the facial expressions in a manner to associate a different speech with a person than the one originally given. Deepfake videos pose a serious threat to legal, political, and social systems as they can destroy the integrity of a person. Research solutions are being designed for the detection of such deepfake content to preserve privacy and combat fake news. This study details the existing deepfake video creation techniques and provides an overview of the deepfake datasets that are publicly available. More importantly, we provide an overview of the deepfake detection methods, along with a discussion on the issues, challenges, and future research directions. The study aims to present an all-inclusive overview of deepfakes by providing insights into the deepfake creation techniques and the latest detection methods, facilitating the development of a robust and effective deepfake detection solution.

Sign in / Sign up

Export Citation Format

Share Document