Filters








19,535 Hits in 4.3 sec

Approximations of Gaussian Process Uncertainties for Visual Recognition Problems [chapter]

Paul Bodesheim, Alexander Freytag, Erik Rodner, Joachim Denzler
2013 Lecture Notes in Computer Science  
Hence, we propose an approximation of the Gaussian process predictive variance leading to rigorous speedups.  ...  Gaussian processes offer the advantage of calculating the classification uncertainty in terms of predictive variance associated with the classification result.  ...  Introduction The Gaussian process framework is a powerful tool for solving regression and classification problems [15] , especially for complex recognition tasks in the visual domain [9, 10, 16, 7] .  ... 
doi:10.1007/978-3-642-38886-6_18 fatcat:kx5afsz63jdotos3qv7rokghni

Active Learning with Gaussian Processes for Object Categorization

Ashish Kapoor, Kristen Grauman, Raquel Urtasun, Trevor Darrell
2007 2007 IEEE 11th International Conference on Computer Vision  
Discriminative methods for visual object category recognition are typically non-probabilistic, predicting class labels but not directly providing an estimate of uncertainty.  ...  Gaussian Processes (GPs) are powerful regression techniques with explicit uncertainty models; we show here how Gaussian Processes with covariance functions defined based on a Pyramid Match Kernel (PMK)  ...  While they have been used in [27, 28] for human motion modeling and in [32] for stereo segmentation, we are unaware of any prior work on visual object recognition in a Gaussian Process framework.  ... 
doi:10.1109/iccv.2007.4408844 dblp:conf/iccv/KapoorGUD07 fatcat:bvlmticc6bfe7gz5e4pyjdwqfu

Multimodal Fusion By Adaptive Compensation For Feature Uncertainty With Application To Audiovisual Speech Recognition

Athanassios Katsamanis, Petros Maragos, George Papandreou, Vassilis Pitsikalis
2006 Zenodo  
Publication in the conference proceedings of EUSIPCO, Florence, Italy, 2006  ...  Gowdy for the use of the CUAVE database.  ...  Potamianos for discussions and for providing the initial experimental setup for AV-ASR, I. Kokkinos for visual front-end discussions, K. Murphy for using his HMM toolkit, and J.N.  ... 
doi:10.5281/zenodo.52626 fatcat:6vw4oeehjjbankw62zdg5ozp2q

Adaptive Multimodal Fusion by Uncertainty Compensation with Application to Audio-Visual Speech Recognition [chapter]

George Papandreou, Athanassios Katsamanis, Athanassios Katsamanis, Vassilis Pitsikalis, Petros Maragos
2008 Multimodal Processing and Interaction  
In this context, we propose improved techniques for person-independent visual feature extraction and uncertainty estimation with active appearance models, and also discuss how enhanced audio features along  ...  We show how these ideas can be practically applied for audiovisual speech recognition.  ...  Kokkinos for visual front-end discussions, K. Murphy for making his HMM toolkit publicly available, and J. N. Gowdy for providing the CUAVE database.  ... 
doi:10.1007/978-0-387-76316-3_4 fatcat:l2odqrkgxvaerkc4h2qnry7lyq

Multiple kernel Gaussian process classification for generic 3D object recognition

Erik Rodner, Doaa Hegazy, Joachim Denzler
2010 2010 25th International Conference of Image and Vision Computing New Zealand  
We study the suitability of approximate GP classification methods for such tasks and present and evaluate different image kernel functions for range and colour images.  ...  We present an approach to generic object recognition with range information obtained using a Time-of-Flight camera and colour images from a visual sensor.  ...  Acknowledgments The authors would like to thank Michael Kemmler for valuable discussions about Gaussian Processes and his comments on the paper.  ... 
doi:10.1109/ivcnz.2010.6148815 fatcat:pium6gvkhff5bibxte4vedqfpi

Adaptive Multimodal Fusion by Uncertainty Compensation With Application to Audiovisual Speech Recognition

George Papandreou, Athanassios Katsamanis, Vassilis Pitsikalis, Petros Maragos
2009 IEEE Transactions on Audio, Speech, and Language Processing  
In this context, we propose improved techniques for person-independent visual feature extraction and uncertainty estimation with active appearance models, and also discuss how enhanced audio features along  ...  We show how these ideas can be practically applied for audiovisual speech recognition.  ...  Kokkinos for visual front-end discussions, K. Murphy for making his HMM toolkit publicly available, and J. N. Gowdy for providing the CUAVE database.  ... 
doi:10.1109/tasl.2008.2011515 fatcat:k2pq2cgyefeedghz3qmeqhelg4

Probabilistic Recurrent State-Space Models [article]

Andreas Doerr, Christian Daniel, Martin Schiegg, Duy Nguyen-Tuong, Stefan Schaal, Marc Toussaint, Sebastian Trimpe
2018 arXiv   pre-print
Fully probabilistic SSMs, however, are often found hard to train, even for smaller problems.  ...  To overcome this limitation, we propose a novel model formulation and a scalable training algorithm based on doubly stochastic variational inference and Gaussian processes.  ...  Acknowledgements This research was supported in part by National Science Foundation grants IIS-1205249, IIS-1017134, EECS-0926052, the Office of Naval Research, the Okawa Foundation, and the Max-Planck-Society  ... 
arXiv:1801.10395v2 fatcat:ysvlttbtmbejxmkpormsmmerma

Fuzzy encoding with hybrid pooling for visual dictionary in food recognition

Mohd Norhisham Razali, Noridayu Manshor, Alfian Abdul Halin, Norwati Mustapha, Razali Yaakob
2021 Indonesian Journal of Electrical Engineering and Computer Science  
The current method based on hard assignment and Fisher vector approach to construct visual dictionary have unexpectedly cause errors from the uncertainty problem during visual word assignation.  ...  <span>Tremendous number of f food images in the social media services can be exploited by using food recognition for healthcare benefits and food industry marketing.  ...  However, despite the success of fuzzy encoding approaches, to the best of our knowledge, the problem of uncertainty and plausibility in food recognition domain have yet to be explored.  ... 
doi:10.11591/ijeecs.v21.i1.pp179-195 fatcat:oxsbsbi42jc2hlem2yo6mtgrb4

Sonification of Probabilistic Feedback through Granular Synthesis

J. Williamson, R. Murray-Smith
2005 IEEE Multimedia  
We have applied these techniques to challenging control problems as well as to the sonification of online probabilistic gesture recognition.  ...  We're using these displays in mobile, gestural interfaces where visual display is often impractical.  ...  For example, given a model of the dynamics of a particular interactive system, where there may be both uncertainty in the current state and uncertainty in the model, Monte Carlo sampling can approximate  ... 
doi:10.1109/mmul.2005.37 fatcat:kbn66dpcrzdgxdvlwvfvh6nxpa

Estimating a Sparse Representation of Gaussian Processes Using Global Optimization and the Bayesian Information Criterion

Wilfried Wöber, Georg Novotny, Mohamed Aburaia, Richard Otrebski, Wilfried Kubinger
2018 Zenodo  
Statistical tools such as Bayes filters are used for localization. The mplementation of Gaussian processes in Bayes filters to estimate transition and measurement models were introduced recently.  ...  Based on visual odometry data of a mobile robot, the method was evaluated.  ...  This work tackles this problem by estimating pseudo-data for a sparse representation of a Gaussian process.  ... 
doi:10.5281/zenodo.3402876 fatcat:wiy7lpzvd5gllkwqecmj5fzwba

Uncertainty-Based Rejection in Machine Learning: Implications for Model Development and Interpretability

Marília Barandas, Duarte Folgado, Ricardo Santos, Raquel Simão, Hugo Gamboa
2022 Electronics  
Prior research focused on the development of methods to quantify uncertainty; however, less attention has been given to how to leverage the knowledge of uncertainty in the process of model development.  ...  Uncertainty is present in every single prediction of Machine Learning (ML) models. Uncertainty Quantification (UQ) is arguably relevant, in particular for safety-critical applications.  ...  In this scenario, the open set recognition and out-of-distribution problems are commonly mentioned [22] .  ... 
doi:10.3390/electronics11030396 fatcat:553rkiojfjen7gyzeo66j3cgti

A Survey on Novel Estimation Approach of Motion Controllers for Self-Driving Cars

Vinothkanna R
2021 Journal of Electronics and Informatics  
Recently, the localization uncertainty control will be estimating by Gaussian framework. This estimation suffers from real time constraint distribution for (Global Positioning System) GPS error.  ...  The motion planning framework is one of the challenging tasks in autonomous driving cars. During motion planning, predicting of trajectory is computed by Gaussian propagation.  ...  The author considered Point based Markov Decision Process (QMDP) [13] above mentioned uncertainty problems of AV to make up efficiently during unsafe conditions.  ... 
doi:10.36548/jei.2020.4.003 fatcat:aq64goeqovadjgkrd4scb2voky

Gaussian Processes for Object Categorization

Ashish Kapoor, Kristen Grauman, Raquel Urtasun, Trevor Darrell
2009 International Journal of Computer Vision  
Discriminative methods for visual object category recognition are typically non-probabilistic, predicting class labels but not directly providing an estimate of uncertainty.  ...  Gaussian Processes (GPs) provide a framework for deriving regression techniques with explicit uncertainty models; we show here how Gaussian Processes with covariance functions defined based on a Pyramid  ...  We also wish to thank Manik Varma for his help in obtaining the required data and for many helpful discussions.  ... 
doi:10.1007/s11263-009-0268-3 fatcat:agm3qlh6fzba3jgo5fcfsgk5jm

Visual Tracking Using Particle Filters with Gaussian Process Regression [chapter]

Hongwei Li, Yi Wu, Hanqing Lu
2009 Lecture Notes in Computer Science  
Particle degeneracy is one of the main problems when particle filters are applied to visual tracking.  ...  The main characteristic of the proposed algorithm is that we incorporate particle filters with Gaussian process regression which can learn highly effective proposal distributions for particle filters to  ...  Introduction Visual tracking is currently one of the most actively researched areas of computer vision and pattern recognition.  ... 
doi:10.1007/978-3-540-92957-4_23 fatcat:3vyuibdygvarznevrody6tluqe

Identification of Confinement Regimes in Tokamak Plasmas by Conformal Prediction on a Probabilistic Manifold [chapter]

Geert Verdoolaege, Jesús Vega, Andrea Murari, Guido Van Oost
2012 IFIP Advances in Information and Communication Technology  
Pattern recognition is becoming an increasingly important tool for making inferences from the massive amounts of data produced in magnetic confinement fusion experiments.  ...  The conformal predictor also returns confidence and credibility measures, which are particularly important for interpretation of pattern recognition results in stochastic fusion data.  ...  For illustration purposes, an (approximately) isometric embedding of the Gaussian manifold in three-dimensional Euclidean space is shown in Figure 1a .  ... 
doi:10.1007/978-3-642-33412-2_25 fatcat:l6dfc32chjhujjx46yceyhc26i
« Previous Showing results 1 — 15 out of 19,535 results