Filters








9,620 Hits in 2.8 sec

Deep Residual Flow for Out of Distribution Detection [article]

Ev Zisselman, Aviv Tamar
2020 arXiv   pre-print
We introduce the residual flow, a novel flow architecture that learns the residual distribution from a base Gaussian distribution.  ...  For out of distribution detection in image datasets, our approach provides a principled improvement over the state-of-the-art.  ...  Residual Flow Applied to OOD Detection We now describe an application of the residual flow that extends the Gaussian model of [27] for OOD detection.  ... 
arXiv:2001.05419v3 fatcat:niealvyeerh7ti5v7367pswnxq

Transfer Learning Methods as a New Approach in Computer Vision Tasks with Small Datasets

Andrzej Brodzicki, Michal Piekarski, Dariusz Kucharski, Joanna Jaworek-Korjakowska, Marek Gorgon
2020 Foundations of Computing and Decision Sciences  
Achieved numbers are state-of-the-art results in melanoma thickness prediction, anomaly detection and Clostridium di cile cytotoxicity classification problems.  ...  The deep residual network deals with problems mentioned earlier by using residual blocks, which take advantage of residual mapping to preserve inputs (Fig. 4b) .  ...  Anomaly detection solution Deep convolutional pre-trained neural network VGG-19 was used to detect abnormal situations in multivariate diagnostic signals.  ... 
doi:10.2478/fcds-2020-0010 fatcat:54qxfmohhrcppm3hj2sq4vi55q

Human hippocampus associates information in memory

K. Henke, B. Weber, S. Kneifel, H. G. Wieser, A. Buck
1999 Proceedings of the National Academy of Sciences of the United States of America  
Recent experimental results suggest that the hippocampal contribution to human memory is limited to episodic memory, novelty detection, semantic (deep) processing of information, and spatial memory.  ...  Neither novelty detection nor depth of processing activated the hippocampal formation as much as semantically associating the primarily unrelated words in memory.  ...  K. von Schulthess for the PET infrastructure, T. Berthold and K. Schwedler for help with the data collection, and the referees for their careful reviews.  ... 
doi:10.1073/pnas.96.10.5884 pmid:10318979 pmcid:PMC21955 fatcat:dngq3trkyjfyxfjoox7wi36zhq

Granular Learning with Deep Generative Models using Highly Contaminated Data [article]

John Just
2020 arXiv   pre-print
An approach to utilize recent advances in deep generative models for anomaly detection in a granular (continuous) sense on a real-world image dataset with quality issues is detailed using recent normalizing  ...  Furthermore, downstream classification is shown to be possible and effective via a weakly supervised approach using the log-likelihood output from a normalizing flow model as a training signal for a feature-extracting  ...  Acknowledgements We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan V GPUs used for this research.  ... 
arXiv:2001.04297v1 fatcat:ikb4bdfjlzcifb3gofk7ylvl64

Comparison of novelty detection methods for multispectral images in rover-based planetary exploration missions

Hannah R. Kerner, Kiri L. Wagstaff, Brian D. Bue, Danika F. Wellington, Samantha Jacob, Paul Horton, James F. Bell, Chiman Kwan, Heni Ben Amor
2020 Data mining and knowledge discovery  
that the latter methods are better suited for detecting spectral novelties-i.e., the best method for a given setting depends on the type of novelties that are sought.  ...  Several novelty detection methods have been explored in prior work for three-channel color images and nonimage datasets, but few have considered multispectral or hyperspectral image datasets for the purpose  ...  Figure 1 shows the flow of Mastcam image data once it arrives at the Deep Space Network.  ... 
doi:10.1007/s10618-020-00697-6 fatcat:lipqcsetwrgh3aft2nz7dyg6iq

Latent Space Autoregression for Novelty Detection

Davide Abati, Angelo Porrello, Simone Calderara, Rita Cucchiara
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Novelty detection is commonly referred to as the discrimination of observations that do not conform to a learned model of regularity.  ...  In our proposal, we design a general framework where we equip a deep autoencoder with a parametric density estimator that learns the probability distribution underlying its latent representations through  ...  We gratefully acknowledge Facebook Artificial Intelligence Research and Panasonic Silicon Valley Lab for the donation of GPUs used for this research.  ... 
doi:10.1109/cvpr.2019.00057 dblp:conf/cvpr/AbatiPCC19 fatcat:tvlifcymbfcfvaqfq3l4fanmti

Detection of Small-Sized Insects in Sticky Trapping Images Using Spectral Residual Model and Machine Learning

Wenyong Li, Zhankui Yang, Jiawei Lv, Tengfei Zheng, Ming Li, Chuanheng Sun
2022 Frontiers in Plant Science  
To alleviate this situation, this study proposed an automated detection approach to meet the need for continuous monitoring of pests in greenhouse conditions.  ...  Candidate targets were firstly located using a spectral residual model and then different color features were extracted.  ...  FIGURE 2 | 2 FIGURE 2 | Flow chart of the candidate object location pipeline from source image to detection results.  ... 
doi:10.3389/fpls.2022.915543 pmid:35837447 pmcid:PMC9274131 fatcat:bgqjtjsdbbcg3og4drmd7mmeke

Detection of Rolling-Element Bearing Faults in Non-stationary Quasi-Parallel Machinery Using Residual Analysis Augmented by Neural Networks

Dustin Helm, Markus Timusk
2021 International Journal of Prognostics and Health Management  
The proposed method is directly compared to a typical AANN novelty detection scheme.  ...  This work proposes a methodology for the detection of rolling-element bearing faults in quasi-parallel machinery.  ...  One of the first implementations of an AANN as a novelty detector for fault detection was done by Japkowitz et al.  ... 
doi:10.36001/ijphm.2021.v12i2.2915 fatcat:ry7ljdagireerlwguvszva5aim

FlowFusion: Dynamic Dense RGB-D SLAM Based on Optical Flow [article]

Tianwei Zhang, Huayan Zhang, Yang Li, Yoshihiko Nakamura, Lei Zhang
2020 arXiv   pre-print
Our novelty is using optical flow residuals to highlight the dynamic semantics in the RGB-D point clouds and provide more accurate and efficient dynamic/static segmentation for camera tracking and background  ...  Dynamic environments are challenging for visual SLAM since the moving objects occlude the static environment features and lead to wrong camera motion estimation.  ...  ACKNOWLEDGEMENTS This work was supported by JSPS Grants-in-Aid for Scientific Research (A) 17H06291. We thank Dr. Raluca Scona for opening source the codes of StaticFusion.  ... 
arXiv:2003.05102v1 fatcat:zo4t4edxgbdfzblpllas2vm2sa

Latent Space Autoregression for Novelty Detection [article]

Davide Abati, Angelo Porrello, Simone Calderara, Rita Cucchiara
2019 arXiv   pre-print
Novelty detection is commonly referred to as the discrimination of observations that do not conform to a learned model of regularity.  ...  In our proposal, we design a general framework where we equip a deep autoencoder with a parametric density estimator that learns the probability distribution underlying its latent representations through  ...  We gratefully acknowledge Facebook Artificial Intelligence Research and Panasonic Silicon Valley Lab for the donation of GPUs used for this research.  ... 
arXiv:1807.01653v2 fatcat:oqjbit3ew5efdntiiayy6ietxa

SDCN2: A Shallow Densely connected CNN for multi-purpose image manipulation detection

Gurinder Singh, Puneet Goyal
2022 ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP)  
The proposed model offers overall detection accuracies of 98.34% and 99.22% for BOSSBase and Dresden datasets, respectively for multiple image manipulation detection.  ...  Also, the hierarchical features produced by the densely connected blocks (DCBs) and prediction error features are fused globally for better information flow across the network.  ...  flow by fusing the output feature maps of all DCBs with feature maps of residual layer to simultaneously consider deep features at both local and global levels  ... 
doi:10.1145/3510462 fatcat:7nlndu7wcfaell2pc2gopkutve

A Comprehensive Review of Deep Learning-Based Crack Detection Approaches

Younes Hamishebahar, Hong Guan, Stephen So, Jun Jo
2022 Applied Sciences  
Critical future directions for research are proposed, based on these reviewed studies as well as on trends and developments in areas similar to the crack detection area.  ...  The application of deep architectures inspired by the fields of artificial intelligence and computer vision has made a significant impact on the task of crack detection.  ...  Ref Novelty/Novelties Method [91] The first application of deep learning for the task of crack segmentation, where using ConvNet, the feature extraction is done on raw images.  ... 
doi:10.3390/app12031374 fatcat:jh4nwibv6ja7xo73kif7lqqzuu

Alpha-Net: Architecture, Models, and Applications [article]

Jishan Shaikh, Adya Sharma, Ankit Chouhan, Avinash Mahawar
2020 arXiv   pre-print
Deep learning network training is usually computationally expensive and intuitively complex. We present a novel network architecture for custom training and weight evaluations.  ...  We provided the empirical mathematical formulation of network loss function for more understanding of accuracy estimation and further optimizations.  ...  Authors are also grateful to their fellow mates for healthy competition and criticism.  ... 
arXiv:2007.07221v1 fatcat:jgcqz34o3fazxl5i6k77dt72eq

Minimal Residual Disease Assessment Within the Bone Marrow of Multiple Myeloma: A Review of Caveats, Clinical Significance and Future Perspectives

Alessandra Romano, Giuseppe Alberto Palumbo, Nunziatina Laura Parrinello, Concetta Conticello, Marina Martello, Carolina Terragna
2019 Frontiers in Oncology  
MFC approaches, which have been progressively improved up to the so-called Next Generation Flow (NGF), and NGS, which proved clear advantages over ASO-PCR, can detect very low levels of residual disease  ...  These methods are actually almost superimposable, in terms of MRD detection power, supporting the lack of unanimous preference for either technique on basis of local availability.  ...  GP reviewed novelties about ASO-RT-PCR. NP reviewed novelties about MFC. MM reviewed novelties about NGS and liquid biopsy. CC reviewed clinical utility of MRD detection.  ... 
doi:10.3389/fonc.2019.00699 pmid:31482061 pmcid:PMC6710454 fatcat:vge45zf77zbq7gfz5littmiwlq

Video-based Abnormal Driving Behavior Detection via Deep Learning Fusions

Wei Huang, Xi Liu, Mingyuan Luo, Peng Zhang, Wei Wang, Jin Wanga
2019 IEEE Access  
) are introduced, to fulfill the video-based abnormal driving behavior detection task for the first time.  ...  For the WGRD and AWGRD, they are more sophisticated as the important idea of residual networks with superpositions of previous layers is incorporated.  ...  detection task for the first time.  ... 
doi:10.1109/access.2019.2917213 fatcat:e2iluuevhfe4njnffvrfddiryi
« Previous Showing results 1 — 15 out of 9,620 results