scholarly journals Image Anomaly Detection Using Normal Data Only by Latent Space Resampling

2020 ◽  
Vol 10 (23) ◽  
pp. 8660
Author(s):  
Lu Wang ◽  
Dongkai Zhang ◽  
Jiahao Guo ◽  
Yuexing Han
Keyword(s):  
Anomaly Detection ◽  
Normal Distribution ◽  
Autoregressive Model ◽  
Probability Model ◽  
Latent Space ◽  
Industrial Inspection ◽  
Detection Stage ◽  
Novel Method ◽  
The Difference ◽  
Reconstruction Model

Detecting image anomalies automatically in industrial scenarios can improve economic efficiency, but the scarcity of anomalous samples increases the challenge of the task. Recently, autoencoder has been widely used in image anomaly detection without using anomalous images during training. However, it is hard to determine the proper dimensionality of the latent space, and it often leads to unwanted reconstructions of the anomalous parts. To solve this problem, we propose a novel method based on the autoencoder. In this method, the latent space of the autoencoder is estimated using a discrete probability model. With the estimated probability model, the anomalous components in the latent space can be well excluded and undesirable reconstruction of the anomalous parts can be avoided. Specifically, we first adopt VQ-VAE as the reconstruction model to get a discrete latent space of normal samples. Then, PixelSail, a deep autoregressive model, is used to estimate the probability model of the discrete latent space. In the detection stage, the autoregressive model will determine the parts that deviate from the normal distribution in the input latent space. Then, the deviation code will be resampled from the normal distribution and decoded to yield a restored image, which is closest to the anomaly input. The anomaly is then detected by comparing the difference between the restored image and the anomaly image. Our proposed method is evaluated on the high-resolution industrial inspection image datasets MVTec AD which consist of 15 categories. The results show that the AUROC of the model improves by 15% over autoencoder and also yields competitive performance compared with state-of-the-art methods.

Variable Precision Depth Encoding for 3D Range Geometry Compression

2020 ◽  
Vol 2020 (17) ◽  
pp. 34-1-34-7
Author(s):  
Matthew G. Finley ◽  
Tyler Bell
Keyword(s):  
Normal Distribution ◽  
Arbitrary Distribution ◽  
File Size ◽  
Geometry Compression ◽  
Variable Precision ◽  
Wide Range ◽  
Novel Method ◽  
Encoding Method ◽  
Rgb Image ◽  
Color Channels

This paper presents a novel method for accurately encoding 3D range geometry within the color channels of a 2D RGB image that allows the encoding frequency—and therefore the encoding precision—to be uniquely determined for each coordinate. The proposed method can thus be used to balance between encoding precision and file size by encoding geometry along a normal distribution; encoding more precisely where the density of data is high and less precisely where the density is low. Alternative distributions may be followed to produce encodings optimized for specific applications. In general, the nature of the proposed encoding method is such that the precision of each point can be freely controlled or derived from an arbitrary distribution, ideally enabling this method for use within a wide range of applications.

scholarly journals A flexible framework for anomaly Detection via dimensionality reduction

2021 ◽  
Author(s):  
Alireza Vafaei Sadr ◽  
Bruce A. Bassett ◽  
M. Kunz
Keyword(s):  
Anomaly Detection ◽  
Dimensionality Reduction ◽  
Dimensional Space ◽  
High Dimensions ◽  
Detection Algorithms ◽  
Latent Space ◽  
Wide Range ◽  
Flexible Framework ◽  
Online Anomaly Detection ◽  
Python Package

AbstractAnomaly detection is challenging, especially for large datasets in high dimensions. Here, we explore a general anomaly detection framework based on dimensionality reduction and unsupervised clustering. DRAMA is released as a general python package that implements the general framework with a wide range of built-in options. This approach identifies the primary prototypes in the data with anomalies detected by their large distances from the prototypes, either in the latent space or in the original, high-dimensional space. DRAMA is tested on a wide variety of simulated and real datasets, in up to 3000 dimensions, and is found to be robust and highly competitive with commonly used anomaly detection algorithms, especially in high dimensions. The flexibility of the DRAMA framework allows for significant optimization once some examples of anomalies are available, making it ideal for online anomaly detection, active learning, and highly unbalanced datasets. Besides, DRAMA naturally provides clustering of outliers for subsequent analysis.

scholarly journals Anomaly detection with convolutional Graph Neural Networks

2021 ◽  
Vol 2021 (8) ◽  
Author(s):  
Oliver Atkinson ◽  
Akanksha Bhardwaj ◽  
Christoph Englert ◽  
Vishal S. Ngairangbam ◽  
Michael Spannowsky
Keyword(s):  
Neural Networks ◽  
Anomaly Detection ◽  
Top Quarks ◽  
New Physics ◽  
Broad Applicability ◽  
Latent Space ◽  
Edge Features ◽  
Graph Neural Networks ◽  
Promising Avenue ◽  
Do So

Abstract We devise an autoencoder based strategy to facilitate anomaly detection for boosted jets, employing Graph Neural Networks (GNNs) to do so. To overcome known limitations of GNN autoencoders, we design a symmetric decoder capable of simultaneously reconstructing edge features and node features. Focusing on latent space based discriminators, we find that such setups provide a promising avenue to isolate new physics and competing SM signatures from sensitivity-limiting QCD jet contributions. We demonstrate the flexibility and broad applicability of this approach using examples of W bosons, top quarks, and exotic hadronically-decaying exotic scalar bosons.

scholarly journals Linear traveltime perturbation interpolation: a novel method to compute 3-D traveltimes

2015 ◽  
Vol 203 (1) ◽  
pp. 548-552 ◽  
Author(s):  
Jianzhong Zhang ◽  
Junjie Shi ◽  
Lin-Ping Song ◽  
Hua-wei Zhou
Keyword(s):  
Seismic Waves ◽  
Homogeneous Medium ◽  
Cell Boundary ◽  
Routine Method ◽  
Bilinear Interpolation ◽  
Linear Distribution ◽  
Heterogeneous Model ◽  
Computational Errors ◽  
Novel Method ◽  
The Difference

Abstract The linear traveltime interpolation has been a routine method to compute first arrivals of seismic waves and trace rays in complex media. The method assumes that traveltimes follow a linear distribution on each boundary of cells. The linearity assumption of traveltimes facilitates the numerical implementation but its violation may result in large computational errors. In this paper, we propose a new way to mitigate the potential shortcoming hidden in the linear traveltime interpolation. We use the vertex traveltimes in a calculated cell to introduce an equivalent homogeneous medium that is specific to the cell boundary from a source. Therefore, we can decompose the traveltime at a point on the cell boundary into two parts: (1) a reference traveltime propagating in the equivalent homogeneous medium and (2) a perturbation traveltime that is defined as the difference between the original and reference traveltimes. We now treat that the traveltime perturbation is linear along each boundary of cells instead of the traveltime. With the new assumption, we carry out the bilinear interpolation over traveltime perturbation to complete traveltime computation in a 3-D heterogeneous model. The numerical experiments show that the new method, the linear traveltime perturbation interpolation, is able to achieve much higher accuracy than that based on the linear traveltime interpolation.

scholarly journals Deceptive Jamming Detection for SAR Based on Cross-Track Interferometry

Sensors ◽  
2018 ◽  
Vol 18 (7) ◽  
pp. 2265 ◽  
Author(s):  
Qingqing Feng ◽  
Huaping Xu ◽  
Zhefeng Wu ◽  
Wei Liu
Keyword(s):  
Sar Interferometry ◽  
Synthetic Aperture ◽  
Sar Image ◽  
Frequency Detection ◽  
Interferometric Phase ◽  
Deceptive Jamming ◽  
Jamming Detection ◽  
Novel Method ◽  
The Difference ◽  
Cross Track

Deceptive jamming against synthetic aperture radar (SAR) can create false targets or deceptive scenes in the image effectively. Based on the difference in interferometric phase between the target and deceptive jamming signals, a novel method for detecting deceptive jamming using cross-track interferometry is proposed, where the echoes with deceptive jamming are received by two SAR antennas simultaneously and the false targets are identified through SAR interferometry. Since the derived false phase is close to a constant in interferogram, it is extracted through phase filtering and frequency detection. Finally, the false targets in the SAR image are obtained according to the detected false part in the interferogram. The effectiveness of the proposed method is validated by simulation results based on the TanDEM-X system.

scholarly journals A NOVEL METHOD OF CLOUD AND SHADOW EXTRACTION WITH MULTI-TEMPORAL LANDSAT8 OLI IMAGES

2017 ◽  
Vol XLII-2/W7 ◽  
pp. 1301-1305
Author(s):  
Y. Ni ◽  
G. He ◽  
W. Jiang
Keyword(s):  
Remote Sensing ◽  
Large Change ◽  
Remote Sensing Image ◽  
Image Process ◽  
Actual Application ◽  
Significant Step ◽  
Multi Temporal ◽  
Ground Object ◽  
Novel Method ◽  
The Difference

Cloud and Shadow removal is a significant step in remote sensing image process. As we all know, the ground object coverage type of the same area of the remote sensing image has little change in the short term. But for cloud and shadow coverage areas, the ground object coverage type has large change. Therefore, according to the difference between the two Landsat / OLI images caused by changes in the cover, this paper presents a method of extracting clouds and shadows based on differences in luminance values. This method selects two thresholds for the difference of brightness values, and extracts the clouds and shadows respectively, and validates them with random point method, which can obtain high precision of extracting cloud and shadow and satisfy the actual application needs.

Novel Method of Assessing Practical Intelligence Acquired in Mechatronics Laboratory Classes

2015 ◽  
pp. 895-928
Author(s):  
Zol Bahri Razali
Keyword(s):  
Engineering Education ◽  
Engineering Students ◽  
Explicit Knowledge ◽  
Assessment Instruments ◽  
Practical Intelligence ◽  
Work Related ◽  
Novel Method ◽  
The Difference ◽  
Representative Work ◽  
Method Of Assessment

Practical intelligence is often referred to as the ability of a person to solve practical challenges in a given domain. The lack of practical intelligence may be due to the way in which explicit knowledge is valued and subsequently assessed in engineering education, namely via examinations, tests, laboratory reports, and tutorial exercises. The lack of effective assessments on practical intelligence indicates implicit devaluation, which can significantly impair engineering students' ability to acquire practical intelligence. To solve this problem, the authors propose a new method of assessment for measuring practical intelligence acquired by engineering students after performing engineering laboratory classes. The novices-experts approach is used in designing the assessment instruments, based on the behaviors' of novices/experts observed and novices/experts representative work-related situations. The practical intelligence can be measured by calculating the difference between participants' and the experts' ratings; the closer the novices to experts, the higher the practical intelligence acquired.

scholarly journals Composite clustering normal distribution transform algorithm

2020 ◽  
Vol 17 (3) ◽  
pp. 172988142091214
Author(s):  
Tian Liu ◽  
Jiongzhi Zheng ◽  
Zhenting Wang ◽  
Zhengdong Huang ◽  
Yongfu Chen
Keyword(s):  
Normal Distribution ◽  
Experimental Result ◽  
Geometric Feature ◽  
Matching Accuracy ◽  
Localization And Mapping ◽  
Density Based Clustering ◽  
Scan Registration ◽  
Novel Method ◽  
Value Decomposition

Scan registration is a fundamental step for the simultaneous localization and mapping of mobile robot. The accuracy of scan registration is critical for the quality of mapping and the accuracy of robot navigation. During all of the scan registration methods, normal distribution transform is an efficient and wild-using one. But normal distribution transform will lead to the unreasonable interruption when splitting the grid and can’t express the points’ local geometric feature by prefixed grid. In this article, we propose a novel method, composite clustering normal distribution transform, which comprises the density-based clustering and k-means clustering to aggregate the points with similar local distributing feature. It takes singular value decomposition to judge the suitable degree of one cluster for further division. Meanwhile, to avoid the radiating phenomenon of LIDAR in measuring the points’ distance, we propose a method based on trigonometric to measure the internal distance. The clustering method in composite clustering normal distribution transform could ensure the expression of LIDAR’s local distribution and matching accuracy. The experimental result demonstrates that our method is more accurate and more stable than the normal distribution transform and iterative closest point methods.

scholarly journals Indicadores de desempenho registrados em função do tempo no basquetebol brasileiro

2018 ◽  
Vol 20 (2) ◽  
pp. 172-181
Author(s):  
Yura Yuka Sato dos Santos ◽  
Lucas Antônio Monezi ◽  
Milton Shoiti Misuta ◽  
Luciano Allegretti Mercadante
Keyword(s):  
Performance Analysis ◽  
Normal Distribution ◽  
Professional Basketball ◽  
Technical Indicators ◽  
Point Field ◽  
Emerging Issues ◽  
Kolmogorov Smirnov ◽  
The Difference ◽  
The Moment ◽  
Smirnov Test

Basketball performance analysis using technical indicators dissociated from the moment they occurred in the game seems to no longer respond to emerging issues of the game as it does not identify the periods when a team’s offensive efficiency has increased or decreased. The aim was to characterize and compare the technical indicators in the positive and negative periods and in the whole game of winning and losing teams in men’s professional basketball. Fourteen games of professional men’s teams of the “Novo Basquete Brasil” Championship in the regular 2011/2012 season were filmed and analyzed. The Kolmogorov-Smirnov test was used to verify data normality. The independent T test was used for variables with normal distribution and the Mann-Whitney test for variables that did not present normal distribution, in order to compare teams’ performance. Analysis in the whole game showed that winning teams had significantly higher averages in successful 3-point field goals but in the positive periods, they showed higher averages for successful free throws, successful layups, defensive rebounds and defensive fouls, and in negative periods, losing teams made more defensive and offensive fouls. The teams’ performance in the whole game may not elucidate the determinant indicators for building the difference in the scoreboard. It is suggested that coaches should identify the periods of best and worst teams’ performance in the game and the indicators involved, preparing teams to overcome the negative periods and obtain more positive periods in the game. 

Sign in / Sign up

Export Citation Format

Share Document