Filters








42 Hits in 5.0 sec

Quantile Linear Algorithm for Robust Binarization of Digitalized Letters

M. Ramirez, E. Tapia, M. Block, R. Rojas
2007 Proceedings of the International Conference on Document Analysis and Recognition  
We describe a threshold-based local algorithm for image binarization.  ...  The binarization threshold is calculated using a statistical model of the high energy pixels. Experiments show that this new approach is faster and better than current state-of-the-art algorithms.  ...  Figure 6 . 6 (a) The Original image and its binarized version using (b) Kavallieratou, (c) Niblack and (d) Quantile Linear binarization. The images are only a detail of a whole letter.  ... 
doi:10.1109/icdar.2007.4377097 dblp:conf/icdar/RamirezTBR07 fatcat:vld6m2vyenbbtjm5xd4y4xbwhe

Document image dewarping using robust estimation of curled text lines

A. Ulges, C.H. Lampert, T.M. Breuel
2005 Eighth International Conference on Document Analysis and Recognition (ICDAR'05)  
Digital cameras have become almost ubiquitous and their use for fast and casual capturing of natural images is unchallenged.  ...  This paper presents a new algorithm for removing both perspective and page curl distortion.  ...  We achieve this estimation using the RAST algorithm, a fast and flexible method for robust geometric model fitting.  ... 
doi:10.1109/icdar.2005.90 dblp:conf/icdar/UlgesLB05 fatcat:eqz5aln2z5hrhg7pszuc2kcud4

Supervised machine learning predictive analytics for alumni income

Daniela A. Gomez-Cravioto, Ramon E. Diaz-Ramos, Neil Hernandez-Gress, Jose Luis Preciado, Hector G. Ceballos
2022 Journal of Big Data  
Background This paper explores machine learning algorithms and approaches for predicting alum income to obtain insights on the strongest predictors and a 'high' earners' class.  ...  We conduct an in-depth analysis to determine whether the accuracy of traditional algorithms can be improved with a data science approach.  ...  The advantage of QR over linear regression is that this method is more robust to outliers and more flexible to the linear assumptions.  ... 
doi:10.1186/s40537-022-00559-6 fatcat:pwy65m6ryvawpcf2htapqzoy2q

On the Accuracy of CRNNs for Line-Based OCR: A Multi-Parameter Evaluation [article]

Bernhard Liebl, Manuel Burghardt
2020 arXiv   pre-print
We show ablations for all components of our training pipeline, which relies on the open source framework Calamari.  ...  We discuss the influence of factors such as binarization, input line height, network width, network depth, and other network training parameters such as dropout.  ...  Computations in this paper were performed on the cluster of the Leipzig University Computing Centre.  ... 
arXiv:2008.02777v1 fatcat:2rfhiddpr5erjcavtxzvezawei

An Explainable and Statistically Validated Ensemble Clustering Model Applied to the Identification of Traumatic Brain Injury Subgroups

Dacosta Yeboah, Louis Steinmeister, Daniel B. Hier, Bassam Hadi, Donald C. Wunsch, Gayla R. Olbricht, Tayo Obafemi-Ajayi
2020 IEEE Access  
This framework for ensemble cluster analysis fully integrates statistical methods at several stages of analysis to enhance the quality and the explainability of results.  ...  We present a framework for an explainable and statistically validated ensemble clustering model applied to Traumatic Brain Injury (TBI).  ...  , Q3: 75% Quantile.  ... 
doi:10.1109/access.2020.3027453 fatcat:ujo3vzyejje4rjsuzydro66woi

A novel pore extraction method for heterogeneous fingerprint images using Convolutional Neural Networks

Ruggero Donida Labati, Angelo Genovese, Enrique Muñoz, Vincenzo Piuri, Fabio Scotti
2017 Pattern Recognition Letters  
Currently, each type of fingerprint acquisition technique (touch-based, touchless, or latent) requires a different algorithm for pore extraction.  ...  Results show that our method is feasible and achieved satisfactory accuracy for all the types of evaluated images, with a better performance with respect to the compared state-of-the-art methods.  ...  region of I, the quantiles 0.1, 0.5 and 0.9 of the local region of I CNN , and the quantiles 0.1, 0.5 and 0.9 of the local region of I C .  ... 
doi:10.1016/j.patrec.2017.04.001 fatcat:lk77vzvfx5djblxjrx2ttyhwha

Self noise and contrast controlled thinning of gray images

Rabaa Youssef, Sylvie Sevestre-Ghalila, Anne Ricordeau, Amel Benazza
2016 Pattern Recognition  
A unitary hypothesis test based on the minimum test statistic is used for the elimination of peaks and noise related extremities, while a fusion of multiple tests is required for the insignificant crest  ...  This leads to a local adjustment and a standardization of the parametric thinning process that depends only on the chosen significance level of the test.  ...  We would like to thank 695 Professor Christine Chappard for providing project image. We also want to thank Professor Dhafer Maalouche for his useful comments and suggestions.  ... 
doi:10.1016/j.patcog.2016.03.033 fatcat:caqvnh5i65aard7dlqowast6fq

Learning Capacity in Simulated Virtual Neurological Procedures

Mattia Samuel Mancosu, Silvester Czanner, Martin Punter
2020 Journal of WSCG  
ACKNOWLEDGMENTS The authors acknowledge the support of the NSERC/Creaform Industrial Research Chair on 3-D Scanning for conducting the work presented in this paper.  ...  ACKNOWLEDGEMENTS The authors would like to thank Oana Rotaru-Orhei for her comments and the three anonymous reviewers for their insightful suggestions.  ...  Especially relevant for the algorithm is the fast and robust evaluation of the empty set and identical set decision algorithms.  ... 
doi:10.24132/csrn.2020.3001.13 fatcat:uytlm7nytrhmnk553ellfhl54a

Identifying gene expression programs of cell-type identity and cellular activity with single-cell RNA-Seq

Dylan Kotliar, Adrian Veres, M Aurel Nagy, Shervin Tabrizi, Eran Hodis, Douglas A Melton, Pardis C Sabeti
2019 eLife  
of cells and tissues.  ...  Identifying gene expression programs underlying both cell-type identity and cellular activities (e.g. life-cycle processes, responses to environmental cues) is crucial for understanding the organization  ...  Acknowledgements We thank Allon Klein, Samuel Wolock, Aubrey Faust, Chris Edwards, Stephen Schaffner, Eric Lander, the CGTA discussion group, and members of the Sabeti Laboratory for useful discussions  ... 
doi:10.7554/elife.43803 pmid:31282856 pmcid:PMC6639075 fatcat:2yuakbxchfcbbiu6gtusnbxbke

Introduction [chapter]

2016 Music Data Analysis  
The interface between the computer and statistical sciences is increasing, as each discipline seeks to harness the power and resources of the other.  ...  This series aims to foster the integration between the computer sciences and statistical, numerical, and probabilistic methods by publishing a broad range of reference works, textbooks, and handbooks.  ...  Hence robustness against such variations is very im portant for the design of transcription systems.  ... 
doi:10.1201/9781315370996-5 fatcat:avooqogcpnbjngqmzuonil3exq

Understanding Hyperdimensional Computing for Parallel Single-Pass Learning [article]

Tao Yu, Yichi Zhang, Zhiru Zhang, Christopher De Sa
2022 arXiv   pre-print
We propose a new theoretical analysis of the limits of HDC via a consideration of what similarity matrices can be "expressed" by binary vectors, and we show how the limits of HDC can be approached using  ...  We propose a new class of VSAs, finite group VSAs, which surpass the limits of HDC.  ...  step with the (nth) quantile function of a Gaussian so as to map into G.  ... 
arXiv:2202.04805v1 fatcat:tzgc5tnyibhhjeeslncic5cose

Modeling & Analysis

2003 NeuroImage  
We have applied the algorithm to several MRI Data Sets. Despite the diversity of the images the neural network shows good robustness.  ...  These findings are consistent with the results of conventional image analysis (SPM99), and they demonstrate that CPCA is a robust method for the examination of the connectivity within, and the interactions  ...  Fisher's linear discriminant is applied to feature vectors for classification.  ... 
doi:10.1016/s1053-8119(05)70006-9 fatcat:zff2suxcofbxvetfrwfwcxi3zm

Advances in Reliably Evaluating and Improving Adversarial Robustness [article]

Jonas Rauber, Universitaet Tuebingen, Bethge, Matthias (Prof. Dr.)
2021
We help uncover and solve this problem through two new types of attacks immune to gradient masking. Misaligned incentives are another reason for insufficient evaluations.  ...  This dissertation summarizes and discusses advances in three areas of adversarial robustness.  ...  It depends on whether we, for example, aim to recognize the postal code on a letter or the digits on a street sign.  ... 
doi:10.15496/publikation-63213 fatcat:pr4xp7fwbfhjhiuzno4e4iwqgm

Forum Bildverarbeitung 2014

Michael Heizmann, Fernando Puente León
2015 TM. Technisches Messen  
Acknowledgments The first author is funded by the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB). We thank Prof. Beyerer for his valuable input.  ...  Binder+Co AG supplied us with a batch of shards of glass waste that we used as sample.  ...  Xie, " Fast and robust circular object detection with probabilistic pairwise voting", Signal Processing Letters, IEEE, Vol. 18, Nr. 11, S. 639-642, Nov. 2011. 6. P. V. C.  ... 
doi:10.1515/teme-2015-0033 fatcat:nzkj6o5l2zcydezksxkcjnpf7a

On layer-wise representations in deep neural networks [article]

Grégoire Montavon, Technische Universität Berlin, Technische Universität Berlin, Klaus-Robert Müller
2013
In this thesis, we develop a kernel-based analysis for deep networks that quantifies the representation at each layer in terms of noise and dimensionality.  ...  The analysis also reveals the disrupting effect of learning noise, and how it prevents the emergence of highly sophisticated deep models.  ...  Learning Molecular Electronic Properties with Neural Networks Binarized (θ = 25) Binarized (θ = 5) Binarized (θ = 1) Figure 5.6.: Linear PCA in the binarized representation space for various granularities  ... 
doi:10.14279/depositonce-3875 fatcat:kdpcdrjptnbariuyingfuf5tyi
« Previous Showing results 1 — 15 out of 42 results