Filters








6,071 Hits in 6.3 sec

Perceiving and reasoning about liquids using fully convolutional networks

Connor Schenck, Dieter Fox
2017 The international journal of robotics research  
We used fully convolutional neural networks to learn to detect and track liquids across pouring sequences.  ...  Our results show that these networks are able to perceive and reason about liquids, and that integrating temporal information is important to performing such tasks well.  ...  For each image in a sequence, we compute the dense optical flow using the same methodology as for the neural network method.  ... 
doi:10.1177/0278364917734052 fatcat:6kvxbcd7azdwhptekblqws2x64

Perceiving and Reasoning About Liquids Using Fully Convolutional Networks [article]

Conor Schenck, Dieter Fox
2017 arXiv   pre-print
We used fully convolutional neural networks to learn to detect and track liquids across pouring sequences.  ...  Our results show that these networks are able to perceive and reason about liquids, and that integrating temporal information is important to performing such tasks well.  ...  Future work will examine methods for training networks to generalize to liquids across many different environments with many different conditions.  ... 
arXiv:1703.01564v2 fatcat:6uba23yziva4znofmecvjdbd2u

Generating Liquid Simulations with Deformation-aware Neural Networks [article]

Lukas Prantl, Boris Bonev, Nils Thuerey
2019 arXiv   pre-print
Our algorithm captures these complex phenomena in two stages: a first neural network computes a weighting function for a set of pre-computed deformations, while a second network directly generates a deformation  ...  To demonstrate the effectiveness of our approach, we showcase our method with several complex examples of flowing liquids with topology changes.  ...  CONCLUSIONS We have presented a novel method to generate space-time surfaces with deformation-aware neural networks.  ... 
arXiv:1704.07854v4 fatcat:benfmgsv6rf4lclrvceesp4py4

Visual perception of liquids: Insights from deep neural networks

Jan Jaap R. van Assen, Shin'ya Nishida, Roland W. Fleming, Daniele Marinazzo
2020 PLoS Computational Biology  
We analysed the network's features using representational similarity analysis (RSA) and a range of image descriptors (e.g. optic flow, colour saturation, GIST).  ...  Liquids and gels are particularly challenging due to their extreme variability and complex behaviour.  ...  Acknowledgments We thank Peter Battaglia for invaluable discussions and help getting started with deep learning and Next Limit for providing support during the stimuli generation process.  ... 
doi:10.1371/journal.pcbi.1008018 pmid:32813688 fatcat:4ogmq7ytrraotcgmsbvmx7626a

Visual perception of liquids: insights from deep neural networks

Jan Jaap R Van Assen, Shin'ya Nishida, Roland W Fleming
2019 Journal of Vision  
PLOS COMPUTATIONAL BIOLOGY Visual perception of liquids: Insights from deep neural networks PLOS Computational Biology | https://doi.  ...  We analysed the network's features using representational similarity analysis (RSA) and a range of image descriptors (e.g. optic flow, colour saturation, GIST).  ...  Acknowledgments We thank Peter Battaglia for invaluable discussions and help getting started with deep learning and Next Limit for providing support during the stimuli generation process.  ... 
doi:10.1167/19.10.242b fatcat:ny7lnqkmbzfalcntv7u3foeug4

Three-dimensional particle tracking velocimetry using shallow neural network for real-time analysis

Yeonghyeon Gim, Dong Kyu Jang, Dong Kee Sohn, Hyoungsoo Kim, Han Seo Ko
2020 Experiments in Fluids  
In this paper, using a shallow neural network model (SNN), we dramatically decrease the computation time with a high accuracy to successfully reconstruct the three-dimensional particle positions, which  ...  The mapping function becomes more complicated if the number of cameras is increased and there is a liquid-vapor interface, which crucially affect the total computation time.  ...  To enhance the speed of the computation time with a high reliability of particle reconstruction, several algorithms have been suggested, such as neural-network PIV using convolutional neural networks (  ... 
doi:10.1007/s00348-019-2861-8 fatcat:qymipuhehrgylbxo2idlfv2yai

Trajectory Extraction and Deep Features for Classification of Liquid-gas Flow under the Context of Forced Oscillation

Luong Nguyen, Julien Mille, Dominique Li, Donatello Conte, Nicolas Ragot
2020 Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications  
From computer vision point of view, it provides a new dynamic texture dataset with challenging tasks since liquid and gas keep changing constantly and the form of liquid-gas flow is closely related to  ...  We show also that a preprocessing step with difference of Gaussian (DoG) over multiple scales as input of deep neural networks is mandatory to obtain satisfying results, up to 81.39% on the test set.  ...  ACKNOWLEDGEMENTS This work was supported by University of Zhejiang and Haoyi Niu in particular. We gratefully acknowledged the support of his work with the video data used for this research.  ... 
doi:10.5220/0008870700170026 dblp:conf/visapp/NguyenMLCR20 fatcat:v5vrywdlhnbgvmjb2t72jvjxha

Independent component analysis of nanomechanical responses of cantilever arrays

Rick Archibald, Panos Datskos, Gerald Devault, Vincent Lamberti, Nickolay Lavrik, Don Noid, Michael Sepaniak, Pampa Dutta
2007 Analytica Chimica Acta  
This report couples the feature extraction abilities of independent component analysis (ICA) and the classification techniques of neural networks to analyze the signals produced by microcantilever-array-based  ...  The ability to detect and identify chemical and biological elements in air or liquid environments is of far reaching importance.  ...  Archibald would like to thank the Householder fellowship that is supported under the Mathematical, Information, and Computational Sciences Division; Office of Advanced Scientific Computing Research; U.S  ... 
doi:10.1016/j.aca.2006.11.007 pmid:17386591 fatcat:wm2zph5swvc77eyh6bnrau2jvi

Visual Closed-Loop Control for Pouring Liquids [article]

Connor Schenck, Dieter Fox
2017 arXiv   pre-print
We combine this with a simple PID controller to pour specific amounts of liquid, and show that the robot is able to achieve an average 38ml deviation from the target amount.  ...  We propose both a model-based and a model-free method utilizing deep learning for estimating the volume of liquid in a container.  ...  Acknowledgement This work was funded in part by the National Science Foundation under contract numbers NSF-NRI-1525251 and NSF-NRI-1637479.  ... 
arXiv:1610.02610v3 fatcat:26yvlkmchrehphaj2not4i5foa

Application of ANN and PCA to two-phase flow evaluation using radioisotopes

Robert Hanus, Marcin Zych, Leszek Petryka, Dariusz Świsulski, Anna Strzępowicz, P. Dančová
2017 EPJ Web of Conferences  
The article shows how features of the signals in the time and frequency domain can be used to build the artificial neural network (ANN) to recognize the structure of the gas-liquid flow in a horizontal  ...  It was found that the reduction of signals features allows for building a network with better performance.  ...  Generally, computational intelligence methods exploit various features of signals in the time, frequency and state-space domain.  ... 
doi:10.1051/epjconf/201714302033 fatcat:f4rigfqhoveq5ky24kr73hr5wy

Learning Similarity Metrics for Numerical Simulations [article]

Georg Kohl, Kiwon Um, Nils Thuerey
2020 arXiv   pre-print
We propose a neural network-based approach that computes a stable and generalizing metric (LSiM) to compare data from a variety of numerical simulation sources.  ...  We leverage a controllable data generation setup with PDE solvers to create increasingly different outputs from a reference simulation in a controlled environment.  ...  We would like to thank Stephan Rasp for preparing the WeatherBench data and all reviewers for helping to improve this work.  ... 
arXiv:2002.07863v2 fatcat:chi2mlzmqjc2bjczbw3e7nyfae

Massively Parallel Amplitude-Only Fourier Neural Network [article]

Mario Miscuglio, Zibo Hu, Shurui Li, Jonathan George, Roberto Capanna, Philippe M. Bardet, Puneet Gupta, Volker J. Sorger
2020 arXiv   pre-print
Conceptually, the information-flow direction is orthogonal to the two-dimensional programmable-network, which leverages 10^6-parallel channels of display technology, and enables a prototype demonstration  ...  We exemplary realize a convolutional neural network (CNN) performing classification tasks on 2-Megapixel large matrices at 10 kHz rates, which latency-outperforms current GPU and phase-based display technology  ...  This paradigm and hardware implementation of the optical engines for artificial neural networks is a promising alternative to other machine learning architecture since they can avail parallel computing  ... 
arXiv:2008.05853v2 fatcat:jghonjlavna4xa72xvbxtbcjhy

Free-form optimization of nanophotonic devices: from classical methods to deep learning

Juho Park, Sanmun Kim, Daniel Wontae Nam, Haejun Chung, Chan Y. Park, Min Seok Jang
2022 Nanophotonics  
Nanophotonic devices have enabled microscopic control of light with an unprecedented spatial resolution by employing subwavelength optical elements that can strongly interact with incident waves.  ...  However, to date, most nanophotonic devices have been designed based on fixed-shape optical elements, and a large portion of their design potential has remained unexplored.  ...  In a tandem configuration, an inverse neural network is followed by a pre-trained forward neural network, which maps the design parameters to optical response space as schematically shown in Figure 10A  ... 
doi:10.1515/nanoph-2021-0713 fatcat:fwstpus3uffzlnv6ddbqzafczq

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Zainoolabadien Karim, Terence L. van Zyl
2021 Remote Sensing  
We introduce a feature space ensemble family (FeatSpaceEnsNet), an average ensemble family (AvgEnsNet), and a hybrid ensemble family (HybridEnsNet) of TLFE neural networks.  ...  Machine learning models using feature descriptors and non-deep learning classifiers, including a two-layer convolutional neural network (ConvNet2), were used as baselines.  ...  Each of the sub-networks in the pair of networks use an independent neural network with identical architecture and embedding feature space length.  ... 
doi:10.3390/rs13214394 fatcat:ph5hktv4dnhdlgihy2olhsbb3q

Large-Scale Optical Reservoir Computing for Spatiotemporal Chaotic Systems Prediction [article]

Mushegh Rafayelyan, Jonathan Dong, Yongqi Tan, Florent Krzakala, and Sylvain Gigan
2020 arXiv   pre-print
Reservoir computing is a relatively recent computational paradigm that originates from a recurrent neural network, and is known for its wide-range of implementations using different physical technologies  ...  is due to electronic overheads, while the optical part of computation remains fully parallel and independent of the reservoir size.  ...  Sylvain Gigan and Jonathan Dong also acknowledge partial support from H2020 European Research Council (ERC) (Grant 724473).  ... 
arXiv:2001.09131v1 fatcat:hzgfpkpynbcbjdu772jxzsju6e
« Previous Showing results 1 — 15 out of 6,071 results