Filters








174 Hits in 5.2 sec

Feature Analysis of Numerical Calculated Data from Sweep Frequency Analysis (SFRA) Traces Using Self Organizing Maps

Z. H. Bohari, S. Ab Ghani, M. F. Baharom, M. N. M. Nasir, M. H. Jali, Y. H. Md Thayoob
2014 Jurnal Teknologi  
low quantization and topographic error.  ...  Both quantization error and topographic error are at 0.009 In the map units or neurons of the U Matrix ( Figure 5) shown are labeled by their respective substation with alphabetical order from A to M  ... 
doi:10.11113/jt.v67.2762 fatcat:yfrguax4dngz3eac5olzefke5i

Modeling cross-sensory and sensorimotor correlations to detect and localize faults in mobile robots

Zsolt Kira
2007 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems  
Results indicate that the system can detect sensory and motor faults in a mobile robot and identify their cause, without using a priori models of the robot or its fault states.  ...  We learn statistical models of correlations between elements in the hierarchy, in addition to the control signals, and use this to detect and identify changes in the capabilities of the robot.  ...  The network consists of a lattice of nodes or cells, usually configured in a rectangular or hexagonal shape.  ... 
doi:10.1109/iros.2007.4398978 dblp:conf/iros/Kira07 fatcat:v6eudsw3uzadlcganl5pchs5gy

Hexagonal Image Processing in the Context of Machine Learning: Conception of a Biologically Inspired Hexagonal Deep Learning Framework [article]

Tobias Schlosser, Michael Friedrich, Danny Kowerko
2019 arXiv   pre-print
, they can also result in a reduction of trainable parameters as well as result in increased training and test rates.  ...  Inspired by the human visual perception system, hexagonal image processing in the context of machine learning deals with the development of image processing systems that combine the advantages of evolutionary  ...  same number of sampling points, which following Golay [10] result in reduced computation times, less quantization errors, and an increased efficiency in programming.  ... 
arXiv:1911.11251v2 fatcat:umi73lbnzjb5fadm63s2i2cnq4

Adaptive nonlinear manifolds and their applications to pattern recognition

Hujun Yin, Weilin Huang
2010 Information Sciences  
A recent comprehensive review can be found in [66] . The SOM is topology-preserving vector quantization.  ...  Self-organization is a fundamental pattern recognition process, in which intrinsic relationships within the sensory data are learned.  ...  In addition to SOM's objective to minimize the quantization error, the aim is also to maintain constant inter-neuron distances locally.  ... 
doi:10.1016/j.ins.2010.04.004 fatcat:zhsym3p2nvhszhnsc7rh6h5i4u

Penrose Pixels Super-Resolution in the Detector Layout Domain

Moshe Ben-Ezra, Zhouchen Lin, Bennett Wilburn
2007 2007 IEEE 11th International Conference on Computer Vision  
We also present a new variant of the well known error back-projection super-resolution algorithm that makes use of the exact detector model in its back projection operator for better accuracy.  ...  Pixels in our model can vary in shape and size, and there may be gaps between adjacent pixels.  ...  (c) reconstructed image RMS error is 9.4. The results gradually degrade as the quantization error and magnification increase.  ... 
doi:10.1109/iccv.2007.4408888 dblp:conf/iccv/Ben-EzraLW07 fatcat:lbxqleyfzzcufmok3douyqi7le

Theory and applications of neural maps

Thomas Villmann, Udo Seiffert, Axel Wismüller
2004 The European Symposium on Artificial Neural Networks  
In this tutorial paper about neural maps we review the current state in theoretical aspects like mathematical treatment of convergence, ordering and topography, magnification and others.  ...  Introduction Fundamentals and basic notations Neural maps occur in real brains in all sensory modalities as well as motor areas being an important step in information processing.  ...  Thereby, in SOM the apriori given lattice A is usually a possibly high-dimensional rectangular lattice or a grid based on hexagonal structures.  ... 
dblp:conf/esann/VillmannSW04 fatcat:z35oqpqzxfc3bcwueomir7lk5y

Developing dually optimal LCA features in sensory and action spaces for classification

Nikita Wagle, Juyang Weng
2012 2012 IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL)  
Since, Z can be taught to represent a set of trainer specified meanings (e.g., type and location), a DN treats these meanings in a unified way for both detection and recognition for objects in dynamic  ...  In this work, we fill this void. We describe how the Z information enables the features to be more sensitive to trainer specified output meanings (e.g., type and location).  ...  that satisfies the distance sensitivity property, adding top-down input from motor Z in addition to bottom-up input X enables the quantization errors for action-relevant subspace X r to be smaller than  ... 
doi:10.1109/devlrn.2012.6400885 dblp:conf/icdl-epirob/WagleW12 fatcat:4h3taeqtyvgm5j5glissewklte

General Riemannian SOM [article]

Jascha A. Schewtschenko
2015 arXiv   pre-print
A deviation between the numerical and analytic stability limit has been observed for the square and hexagonal Euclidean maps for very small neighbourhoods in the map space as well as agreement in case  ...  We furthermore perform the analytic and numerical analysis of the stability limits of certain (GRi)SOM configurations covering the different possible regular tessellation of the map space in each geometry  ...  This can be easily checked by plotting the logarithm of this error, which has been done in Fig.26 for the hexagonal map.  ... 
arXiv:1505.03917v1 fatcat:tyhrmjkuf5dx3gmjsfc6gamuem

Simulating auditory and visual sensorineural prostheses: a comparative review

L E Hallum, G Dagnelie, G J Suaning, N H Lovell
2007 Journal of Neural Engineering  
experimental work in coming years.  ...  in doing so restore some spatial vision.  ...  For this reason, Hallum et al [75] proposed the hexagonal arrangement of phosphenes in the visual field (and accordingly the hexagonal manufacture of implantable arrays) which makes for 14% higher density  ... 
doi:10.1088/1741-2560/4/1/s08 pmid:17325417 fatcat:hoopnvhmdfe37gskadjydjneom

ViSOM - a novel method for multivariate data projection and structure visualization

Hujun Yin
2002 IEEE Transactions on Neural Networks  
It produces a graded mesh in the data space such that the distances between mapped data points on the map resemble those in the original space, like in the Sammon mapping.  ...  In this paper, a visualization-induced SOM (ViSOM) is proposed to overcome these shortcomings.  ...  Its distribution and convergence properties show that the SOM, when the neighborhood diminishes, is naturally an optimal vector quantizer (VQ) in minimizing the mean-square error between reference vectors  ... 
doi:10.1109/72.977314 pmid:18244423 fatcat:zwv7li2ej5cgvmjqqenn5w3hma

The Self-Organizing Maps: Background, Theories, Extensions and Applications [chapter]

Hujun Yin
2008 Studies in Computational Intelligence  
Yin to serve as an updated, extended tutorial, a review, as well as a reference for advanced topics in the subject.  ...  The SOM was proposed to model the sensory-to-cortex mapping thus the unsupervised associative memory mechanism. Such a mechanism is also related to vector quantization (VQ) [63] in coding terms.  ...  The results show that the resulting quantization error is smaller. Such a SOM-based VQ has also been applied to video compression [2, 22] for improved performance at low bit rates.  ... 
doi:10.1007/978-3-540-78293-3_17 fatcat:5fxzbyyoofd2ppfrnncyqfji6m

Quantizing Euclidean motions via double-coset decomposition [article]

Christian Wülker, Gregory S. Chirikjian
2018 arXiv   pre-print
Concepts from mathematical crystallography and group theory are used here to quantize the group of rigid-body motions, resulting in a "motion alphabet" with which to express robot motion primitives.  ...  More specifically, a very efficient, equivolumetric quantization of spatial motion can be defined using the group-theoretic concept of a double-coset decomposition of the form ΓSE(3) / Δ, where Γ is a  ...  We chose g in this particular way so as to minimize the round-off error in the quantization scheme.  ... 
arXiv:1811.11640v1 fatcat:ieg2dmdmojgqtlrdhogj4cohey

Learning to track colored objects with log-polar vision

Giorgio Metta, Antonios Gasteratos, Gulio Sandini
2004 Mechatronics (Oxford)  
Most important, the parameters of the controllers are learnt on-line in a self-supervised fashion.  ...  Color information is used to determine the position of the object of interest in the image plane and, consequently, to track it during its motion.  ...  of a bin is comparable to the quantization step in the color space.  ... 
doi:10.1016/j.mechatronics.2004.05.003 fatcat:j4n4ubrglrhapogc64rn2y3gvm

Quantizing Euclidean Motions via Double-Coset Decomposition

Christian Wülker, Sipu Ruan, Gregory S. Chirikjian
2019 Research  
Concepts from mathematical crystallography and group theory are used here to quantize the group of rigid-body motions, resulting in a "motion alphabet" with which robot motion primitives are expressed.  ...  Moreover, the general "signals to symbols" problem in artificial intelligence is cast in this framework for robots moving continuously in the world.  ...  We chose in this particular way so as to minimize the round-off error in the quantization scheme.  ... 
doi:10.34133/2019/1608396 pmid:32043079 pmcid:PMC7006946 fatcat:5dxkvrfzbja63jigrpkd3pifi4

Simulating prosthetic vision: I. Visual models of phosphenes

Spencer C. Chen, Gregg J. Suaning, John W. Morley, Nigel H. Lovell
2009 Vision Research  
A standardized simulation and reporting framework is proposed so that future simulations may be configured to be more realistic to the experience of implant recipients, and the simulation parameters from  ...  In prosthetic vision a visual scene is composed of relatively large, isolated, spots of light so-called "phosphenes", very much like a magnified pictorial print.  ...  In the literature, a square phosphene receptive field is adopted for a square lattice, and a circular or hexagonal phosphene receptive field for hexagonal lattices.  ... 
doi:10.1016/j.visres.2009.02.003 pmid:19504749 fatcat:d3my7gymuffeljuqudz35lnkf4
« Previous Showing results 1 — 15 out of 174 results