scholarly journals Deepfake detection by human crowds, machines, and machine-informed crowds

2021 ◽  
Vol 119 (1) ◽  
pp. e2110013119
Author(s):  
Matthew Groh ◽  
Ziv Epstein ◽  
Chaz Firestone ◽  
Rosalind Picard
Keyword(s):  
Computer Vision ◽  
Visual Processing ◽  
Detection Performance ◽  
Cognitive Capacities ◽  
Detection Model ◽  
Machine Performance ◽  
Recent Emergence ◽  
Model Predictions ◽  
Human Participants ◽  
The Impact

The recent emergence of machine-manipulated media raises an important societal question: How can we know whether a video that we watch is real or fake? In two online studies with 15,016 participants, we present authentic videos and deepfakes and ask participants to identify which is which. We compare the performance of ordinary human observers with the leading computer vision deepfake detection model and find them similarly accurate, while making different kinds of mistakes. Together, participants with access to the model’s prediction are more accurate than either alone, but inaccurate model predictions often decrease participants’ accuracy. To probe the relative strengths and weaknesses of humans and machines as detectors of deepfakes, we examine human and machine performance across video-level features, and we evaluate the impact of preregistered randomized interventions on deepfake detection. We find that manipulations designed to disrupt visual processing of faces hinder human participants’ performance while mostly not affecting the model’s performance, suggesting a role for specialized cognitive capacities in explaining human deepfake detection performance.

scholarly journals A Novel Detection Performance Modular Evaluation Metric of Space-based Infrared System

2021 ◽  
Author(s):  
Xiaoxuan Zhou ◽  
Xinyue Ni ◽  
Jingwen Zhang ◽  
Dongshan Weng ◽  
Zhuoyue Hu ◽  
...  
Keyword(s):  
Detection System ◽  
Detection Performance ◽  
Parameter Design ◽  
Infrared Detection ◽  
System Parameters ◽  
Detection Model ◽  
Fifth Generation ◽  
The Earth ◽  
Evaluation Metric ◽  
The Impact

Abstract In order to reflect the space-based full chain information of the detection process comprehensively and objectively, we proposed a novel modular evaluation metric to discuss the target, background and system independently. It takes the equivalent radiation intensity as the parameter, which can evaluate the detection performance of the system quantitatively. In this paper, taking the fifth-generation American stealth fighter F22 as an example, the mathematical detection model of the space-based infrared system to aircraft targets in the Earth background is described. A modular evaluation metric is proposed. The simulation analyzes the impact of different detection scenes and system parameters on system equivalent irradiance. Furthermore, recommendations for the optimization of the detection system are given. The research results provide a new idea for the analysis of the detection performance of highly maneuverable targets under dynamic backgrounds and have guiding significance for the performance evaluation and parameter design of the infrared detection system.

scholarly journals End-To-End Computer Vision Framework: An Open-Source Platform for Research and Education

Sensors ◽  
2021 ◽  
Vol 21 (11) ◽  
pp. 3691
Author(s):  
Ciprian Orhei ◽  
Silviu Vert ◽  
Muguras Mocofan ◽  
Radu Vasiu
Keyword(s):  
Machine Learning ◽  
Image Processing ◽  
Computer Vision ◽  
Open Source ◽  
Visual Processing ◽  
Research Field ◽  
Learning Models ◽  
Research Activity ◽  
End To End ◽  
Machine Learning Models

Computer Vision is a cross-research field with the main purpose of understanding the surrounding environment as closely as possible to human perception. The image processing systems is continuously growing and expanding into more complex systems, usually tailored to the certain needs or applications it may serve. To better serve this purpose, research on the architecture and design of such systems is also important. We present the End-to-End Computer Vision Framework, an open-source solution that aims to support researchers and teachers within the image processing vast field. The framework has incorporated Computer Vision features and Machine Learning models that researchers can use. In the continuous need to add new Computer Vision algorithms for a day-to-day research activity, our proposed framework has an advantage given by the configurable and scalar architecture. Even if the main focus of the framework is on the Computer Vision processing pipeline, the framework offers solutions to incorporate even more complex activities, such as training Machine Learning models. EECVF aims to become a useful tool for learning activities in the Computer Vision field, as it allows the learner and the teacher to handle only the topics at hand, and not the interconnection necessary for visual processing flow.

Impact of random soil properties on stress–strain response

2004 ◽  
Vol 41 (2) ◽  
pp. 351-355 ◽  
Author(s):  
Dieter Stolle ◽  
Peijun Guo ◽  
Gabriel Sedran
Keyword(s):  
Soil Properties ◽  
Constitutive Modelling ◽  
Soil Parameters ◽  
Strain Response ◽  
Constitutive Behaviour ◽  
Stress Strain ◽  
Deterministic Analysis ◽  
Model Predictions ◽  
Random Variability ◽  
The Impact

This paper analyzes the impact of natural random variation of soil properties on the constitutive modelling of geomaterial behaviour. A theoretical framework for accommodating variation in soil properties is presented. The framework is then used to examine the consequence of parameter variability on stress–strain relations. An important observation is that average soil parameters from a series of tests on small specimens, in which density of the specimens varies randomly, do not necessarily reflect the average constitutive behaviour of soil. Model predictions are shown to be consistent with the experimental data.Key words: random variability, deterministic analysis, soil parameters, constitutive model.

scholarly journals Computer Programming E-Learners’ Personality Traits, Self-Reported Cognitive Abilities, and Learning Motivating Factors

2021 ◽  
Vol 11 (9) ◽  
pp. 1205
Author(s):  
Aiste Dirzyte ◽  
Aivaras Vijaikis ◽  
Aidas Perminas ◽  
Romualda Rimasiute-Knabikiene ◽  
Lukas Kaminskis ◽  
...  
Keyword(s):  
Personality Traits ◽  
Visual Processing ◽  
Cognitive Abilities ◽  
Computer Programming ◽  
Social Pressure ◽  
Negative Emotionality ◽  
Future Research ◽  
Motivating Factors ◽  
Individual Attitude ◽  
The Impact

Educational systems around the world encourage students to engage in programming activities, but programming learning is one of the most challenging learning tasks. Thus, it was significant to explore the factors related to programming learning. This study aimed to identify computer programming e-learners’ personality traits, self-reported cognitive abilities and learning motivating factors in comparison with other e-learners. We applied a learning motivating factors questionnaire, the Big Five Inventory—2, and the SRMCA instruments. The sample consisted of 444 e-learners, including 189 computer programming e-learners, the mean age was 25.19 years. It was found that computer programming e-learners demonstrated significantly lower scores of extraversion, and significantly lower scores of motivating factors of individual attitude and expectation, reward and recognition, and punishment. No significant differences were found in the scores of self-reported cognitive abilities between the groups. In the group of computer programming e-learners, extraversion was a significant predictor of individual attitude and expectation; conscientiousness and extraversion were significant predictors of challenging goals; extraversion and agreeableness were significant predictors of clear direction; open-mindedness was a significant predictor of a diminished motivating factor of punishment; negative emotionality was a significant predictor of social pressure and competition; comprehension-knowledge was a significant predictor of individual attitude and expectation; fluid reasoning and comprehension-knowledge were significant predictors of challenging goals; comprehension-knowledge was a significant predictor of clear direction; and visual processing was a significant predictor of social pressure and competition. The SEM analysis demonstrated that personality traits (namely, extraversion, conscientiousness, and reverted negative emotionality) statistically significantly predict learning motivating factors (namely, individual attitude and expectation, and clear direction), but the impact of self-reported cognitive abilities in the model was negligible in both groups of participants and non-participants of e-learning based computer programming courses; χ² (34) = 51.992, p = 0.025; CFI = 0.982; TLI = 0.970; NFI = 0.950; RMSEA = 0.051 [0.019–0.078]; SRMR = 0.038. However, as this study applied self-reported measures, we strongly suggest applying neurocognitive methods in future research.

scholarly journals Exploring the associations between transcript levels and fluxes in constraint-based models of metabolism

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Neeraj Sinha ◽  
Evert M. van Schothorst ◽  
Guido J. E. J. Hooiveld ◽  
Jaap Keijer ◽  
Vitor A. P. Martins dos Santos ◽  
...  
Keyword(s):  
Metabolic Flux ◽  
Transcript Levels ◽  
Proportionality Constant ◽  
Model Predictions ◽  
Physiological Information ◽  
Highly Expressed Genes ◽  
Transcriptomics Data ◽  
Genome Scale ◽  
The Impact ◽  
Constraint Based Models

Abstract Background Several computational methods have been developed that integrate transcriptomics data with genome-scale metabolic reconstructions to increase accuracy of inferences of intracellular metabolic flux distributions. Even though existing methods use transcript abundances as a proxy for enzyme activity, each method uses a different hypothesis and assumptions. Most methods implicitly assume a proportionality between transcript levels and flux through the corresponding function, although these proportionality constant(s) are often not explicitly mentioned nor discussed in any of the published methods. E-Flux is one such method and, in this algorithm, flux bounds are related to expression data, so that reactions associated with highly expressed genes are allowed to carry higher flux values. Results Here, we extended E-Flux and systematically evaluated the impact of an assumed proportionality constant on model predictions. We used data from published experiments with Escherichia coli and Saccharomyces cerevisiae and we compared the predictions of the algorithm to measured extracellular and intracellular fluxes. Conclusion We showed that detailed modelling using a proportionality constant can greatly impact the outcome of the analysis. This increases accuracy and allows for extraction of better physiological information.

scholarly journals Integrating morphologic and molecular histopathological features through whole slide image registration and deep learning

2022 ◽  
Author(s):  
Kevin Faust ◽  
Michael K Lee ◽  
Anglin Dent ◽  
Clare Fiala ◽  
Alessia Portante ◽  
...  
Keyword(s):  
Computer Vision ◽  
Deep Learning ◽  
Digital Pathology ◽  
Scale Invariant ◽  
Idh1 R132h ◽  
Recent Emergence ◽  
Slide Image ◽  
Discovery Potential ◽  
Diffuse Gliomas ◽  
And Shifts

Abstract Background Modern molecular pathology workflows in neuro-oncology heavily rely on the integration of morphologic and immunohistochemical patterns for analysis, classification, and prognostication. However, despite the recent emergence of digital pathology platforms and artificial intelligence-driven computational image analysis tools, automating the integration of histomorphologic information found across these multiple studies is challenged by large files sizes of whole slide images (WSIs) and shifts/rotations in tissue sections introduced during slide preparation. Methods To address this, we develop a workflow that couples different computer vision tools including scale-invariant feature transform (SIFT) and deep learning to efficiently align and integrate histopathological information found across multiple independent studies. We highlight the utility and automation potential of this workflow in the molecular subclassification and discovery of previously unappreciated spatial patterns in diffuse gliomas. Results First, we show how a SIFT-driven computer vision workflow was effective at automated WSI alignment in a cohort of 107 randomly selected surgical neuropathology cases (97/107 (91%) showing appropriate matches, AUC = 0.96). This alignment allows our AI-driven diagnostic workflow to not only differentiate different brain tumor types, but also integrate and carry out molecular subclassification of diffuse gliomas using relevant immunohistochemical biomarkers (IDH1-R132H, ATRX). To highlight the discovery potential of this workflow, we also examined spatial distributions of tumors showing heterogenous expression of the proliferation marker MIB1 and Olig2. This analysis helped uncovered an interesting and unappreciated association of Olig2 positive and proliferative areas in some gliomas (r = 0.62). Conclusion This efficient neuropathologist-inspired workflow provides a generalizable approach to help automate a variety of advanced immunohistochemically compatible diagnostic and discovery exercises in surgical neuropathology and neuro-oncology.

scholarly journals Mitochondrial DNAs provide insight into trypanosome phylogeny and molecular evolution

2020 ◽  
Author(s):  
Christopher Kay ◽  
Tom A Williams ◽  
Wendy Gibson
Keyword(s):  
Molecular Evolution ◽  
Trypanosoma Brucei ◽  
Rna Editing ◽  
Phylogenetic Trees ◽  
Nucleotide Composition ◽  
Sub Saharan Africa ◽  
African Trypanosomes ◽  
Animal Pathogens ◽  
Recent Emergence ◽  
The Impact

Abstract Background: Trypanosomes are single-celled eukaryotic parasites characterised by the unique biology of their mitochondrial DNA (mtDNA). African livestock trypanosomes impose a major burden on agriculture across sub-Saharan Africa, but are poorly understood compared to those that cause sleeping sickness and Chagas disease in humans. Here we explore the potential of trypanosome mtDNA to study the evolutionary history of trypanosomes and the molecular evolution of their mtDNAs.Results: We used long-read sequencing to completely assemble mtDNAs from four previously uncharacterized African trypanosomes, and leveraged these assemblies to scaffold and assemble a further 103 trypanosome mtDNAs from published short-read data. While synteny was largely conserved, there were repeated, independent losses of Complex I genes. Comparison of edited and non-edited genes revealed the impact of RNA editing on nucleotide composition, with non-edited genes approaching the limits of GC loss. African tsetse-transmitted trypanosomes showed high levels of RNA editing compared to other trypanosomes. Whole mtDNA coding regions were used to construct time-resolved phylogenetic trees, revealing deep divergence events among isolates of the pathogens Trypanosoma brucei and T. congolense .Conclusions: Our mtDNA data represents a new resource for experimental and evolutionary analyses of trypanosome phylogeny, molecular evolution and function. Molecular clock analyses yielded a timescale for trypanosome evolution congruent with major biogeographical events in Africa and revealed the recent emergence of Trypanosoma brucei gambiense and T. equiperdum , major human and animal pathogens.

scholarly journals Decreased P2 Waveform Reflects Impaired Brain Executive Function Induced by 12 h of Low Homeostatic Sleep Pressure: Evidence From an Event-Related Potential Study

2021 ◽  
Vol 15 ◽  
Author(s):  
Lingjing Zeng ◽  
Haijing Wu ◽  
Jialu Li ◽  
Haiteng Wang ◽  
Songyue Xie ◽  
...  
Keyword(s):  
Executive Function ◽  
Response Time ◽  
Visual Processing ◽  
Repeated Measures ◽  
Young Male ◽  
Event Related Potential ◽  
Peak Latency ◽  
Peak Amplitude ◽  
Eeg Data ◽  
The Impact

Homeostatic sleep pressure can cause cognitive impairment, in which executive function is the most affected. Previous studies have mainly focused on high homeostatic sleep pressure (long-term sleep deprivation); thus, there is still little related neuro-psycho-physiological evidence based on low homeostatic sleep pressure (12 h of continuous wakefulness) that affects executive function. This study aimed to investigate the impact of lower homeostatic sleep pressure on executive function. Our study included 14 healthy young male participants tested using the Go/NoGo task in normal resting wakefulness (10:00 am) and after low homeostatic sleep pressure (10:00 pm). Behavioral data (response time and accuracy) were collected, and electroencephalogram (EEG) data were recorded simultaneously, using repeated measures analysis of variance for data analysis. Compared with resting wakefulness, the participants’ response time to the Go stimulus was shortened after low homeostatic sleep pressure, and the correct response rate was reduced. Furthermore, the peak amplitude of Go–P2 decreased significantly, and the peak latency did not change significantly. For NoGo stimulation, the peak amplitude of NoGo–P2 decreased significantly (p < 0.05), and the peak latency was significantly extended (p < 0.05). Thus, the P2 wave is likely related to the attention and visual processing and reflects the early judgment of the perceptual process. Therefore, the peak amplitude of Go–P2 and NoGo–P2 decreased, whereas the peak latency of NoGo–P2 increased, indicating that executive function is impaired after low homeostatic sleep pressure. This study has shown that the P2 wave is a sensitive indicator that reflects the effects of low homeostatic sleep pressure on executive function, and that it is also an important window to observe the effect of homeostatic sleep pressure and circadian rhythm on cognitive function.

A Comparison of Orthogonal Transforms in Engineering Computer Vision

1990 ◽  
Vol 204 (5) ◽  
pp. 321-327 ◽  
Author(s):  
G A H Al-Kindi ◽  
R M Baul ◽  
K F Gill
Keyword(s):  
Computer Vision ◽  
Machine Vision ◽  
Visual Inspection ◽  
Computational Time ◽  
Transform Coefficients ◽  
Orthogonal Transforms ◽  
The Impact

A comparison of a number of commonly used orthogonal transforms, when applied to the recognition and visual inspection of engineering components, has been made. The impact on the performance and computational time for the machine vision process due to varying numbers of transform coefficients is assessed.

scholarly journals Feature Diagnosticity Affects Representations of Novel and Familiar Objects

2014 ◽  
Vol 26 (12) ◽  
pp. 2735-2749 ◽  
Author(s):  
Nina S. Hsu ◽  
Margaret L. Schlichting ◽  
Sharon L. Thompson-Schill
Keyword(s):  
Fruits And Vegetables ◽  
Temporal Cortex ◽  
Familiar Object ◽  
Behavioral Effects ◽  
Diagnostic Features ◽  
Novel Objects ◽  
Human Participants ◽  
Ventral Temporal Cortex ◽  
Concept Representations ◽  
The Impact

Many features can describe a concept, but only some features define a concept in that they enable discrimination of items that are instances of a concept from (similar) items that are not. We refer to this property of some features as feature diagnosticity. Previous work has described the behavioral effects of feature diagnosticity, but there has been little work on explaining why and how these effects arise. In this study, we aimed to understand the impact of feature diagnosticity on concept representations across two complementary experiments. In Experiment 1, we manipulated the diagnosticity of one feature, color, for a set of novel objects that human participants learned over the course of 1 week. We report behavioral and neural evidence that diagnostic features are likely to be automatically recruited during remembering. Specifically, individuals activated color-selective regions of ventral temporal cortex (specifically, left fusiform gyrus and left inferior temporal gyrus) when thinking about the novel objects, although color information was never explicitly probed during the task. Moreover, multiple behavioral and neural measures of the effects of feature diagnosticity were correlated across participants. In Experiment 2, we examined relative color association in familiar object categories, which varied in feature diagnosticity (fruits and vegetables, household items). Taken together, these results offer novel insights into the neural mechanisms underlying concept representations by demonstrating that automatic recruitment of diagnostic information gives rise to behavioral effects of feature diagnosticity.

Sign in / Sign up

Export Citation Format

Share Document