Filters








6,783 Hits in 4.5 sec

Learning with limited and noisy tagging

Yingming Li, Zhongang Qi, Zhongfei (Mark) Zhang, Ming Yang
2013 Proceedings of the 21st ACM international conference on Multimedia - MM '13  
a semi-parametric regularization and takes advantage of the multi-label constraints into the optimization.  ...  While SpSVM-MC is a general method for learning with limited and noisy tagging, in the evaluations we focus on the specific application of noisy image tagging with limited labeled training samples on a  ...  This work is supported in part by the National Basic Research Program of China (2012CB316400), Zhejiang University -Alibaba Financial Joint lab, and Zhejiang Provincial Engineering Center on Media Data  ... 
doi:10.1145/2502081.2502111 dblp:conf/mm/LiQZY13 fatcat:qpe7mt524ndbdjlqa5y3vhqjtm

Adaptive Early-Learning Correction for Segmentation from Noisy Annotations [article]

Sheng Liu, Kangning Liu, Weicheng Zhu, Yiqiu Shen, Carlos Fernandez-Granda
2022 arXiv   pre-print
Deep learning in the presence of noisy annotations has been studied extensively in classification, but much less in segmentation tasks.  ...  It also provides robustness to realistic noisy annotations present in weakly-supervised semantic segmentation, achieving state-of-the-art results on PASCAL VOC 2012.  ...  Related work Classification from noisy labels. Early learning and memorization were first discovered in image classification from noisy labels [33] .  ... 
arXiv:2110.03740v2 fatcat:z7ng6gfwtnc5xcmghn75lhub5a

Semi-supervised learning for photometric supernova classification★

Joseph W. Richards, Darren Homrighausen, Peter E. Freeman, Chad M. Schafer, Dovi Poznanski
2011 Monthly notices of the Royal Astronomical Society  
Our approach is to first use the nonlinear dimension reduction technique diffusion map to detect structure in a database of supernova light curves and subsequently employ random forest classification on  ...  As supernova numbers increase, our semi-supervised method efficiently utilizes this information to improve classification, a property not enjoyed by template based methods.  ...  AC K N OW L E D G M E N T S JWR acknowledges the generous support of a Cyber-Enabled Discovery and Innovation (CDI) grant (0941742) from the National Science Foundation.  ... 
doi:10.1111/j.1365-2966.2011.19768.x fatcat:5bsefsw2ireylmt6kkygdzwera

Graphical Models and Exponential Families [article]

Dan Geiger, Christopher Meek
2013 arXiv   pre-print
We provide a classification of graphical models according to their representation as subfamilies of exponential families.  ...  local distributions, are curved exponential families (CEFs) and graphical models with hidden variables are stratified exponential families (SEFs).  ...  stratifica tion, and to Steffen Lauritzen for guiding us through the mysteries of exponential families.  ... 
arXiv:1301.7376v1 fatcat:kuraib3zubhm7lw6xi2yorcuca

Parametric models for facial features segmentation

Z. Hammal, N. Eveno, A. Caplier, Py. Coulon
2006 Signal Processing  
A specific parametric model is defined for each deformable feature, each model being able to take into account all the possible deformations.  ...  This facial features segmentation is the first step of a set of multi-media applications.  ...  To summarize, we propose the following parametric model well suited to the multiple possible shapes of eyes and eyebrows (see Figure 4 ): • For each eye: a circle for the iris (it could be a semi-circle  ... 
doi:10.1016/j.sigpro.2005.06.006 fatcat:c63vhdgx3fae7bzrhveg4d25qe

Application Of Fuzzy System In Segmentation Of MRI Brain Tumor [article]

Mrigank Rajya, Sonal Rewri, Swati Sheoran
2010 arXiv   pre-print
Level set evolution combining global smoothness with the flexibility of topology changes offers significant advantages over the conventional statistical classification followed by mathematical morphology  ...  Segmentation of images holds an important position in the area of image processing.  ...  There are two forms of deformable models. In the parametric form, also referred to as snakes, an explicit parametric representation of the curve is used.  ... 
arXiv:1005.4292v1 fatcat:kzsbkmo2qje6dcsdspjfyj7lvu

Experiments on Sensitivity of Template Matching for Lung Nodule Detection in Low Dose CT Scans

Shireen Y Elhabian, Hossam Abd el Munim, Salwa Elshazly, AlyA. Farag, Mohamed Aboelghar
2007 2007 IEEE International Symposium on Signal Processing and Information Technology  
purpose of detection, we used deformable circular, and classification.  ...  real nodules of different truth and is described by hit rate curves indicating the proba-978-1 -4244-1 835-0/07/$25.00 ©2007 IEEE  ...  In case of noisy as shown in low dose chest CT scans.  ... 
doi:10.1109/isspit.2007.4458213 fatcat:grg3snimwzecldtk6s4fklfqdy

"Noisy beets": impact of phenotyping errors on genomic predictions for binary traits in Beta vulgaris

Filippo Biscarini, Nelson Nazzicari, Chiara Broccanello, Piergiorgio Stevanato, Simone Marini
2016 Plant Methods  
In particular, KNN outperformed all other classifiers with AUC (area under the ROC curve) higher than 0.95 up to 20 % noisy labels. The runner-up method, RF, had an AUC of 0.941 with 20 % noise.  ...  Conclusions: Local classification methods-KNN and RF-showed higher tolerance to noisy labels compared to methods that leverage global data properties-LR and the two SVM models.  ...  Simone Marini is an International Research Fellow of the Japan Society for the Promotion of Science.  ... 
doi:10.1186/s13007-016-0136-4 pmid:27437026 pmcid:PMC4949885 fatcat:7p5x2xxwwvbohhqackjql2orju

A Semi-parametric Technique for the Quantitative Analysis of Dynamic Contrast-enhanced MR Images Based on Bayesian P-splines [article]

Volker J. Schmid, Brandon Whitcher, Anwar R. Padhani, Guang-Zhong Yang
2008 arXiv   pre-print
To overcome this problems, this paper proposes a semi-parametric penalized spline smoothing approach, with which the AIF is convolved with a set of B-splines to produce a design matrix using locally adaptive  ...  Quantitative analysis of DCE-MRI typically involves the convolution of an arterial input function (AIF) with a nonlinear pharmacokinetic model of the contrast agent concentration.  ...  We are grateful to David Buckley at Imaging Science and Biomedical Engineering, University of Manchester, UK for providing the simulated data.  ... 
arXiv:0801.4065v1 fatcat:6n4ii2qbu5cnditxubnedvhaoq

Temporal Ensembling for Semi-Supervised Learning [article]

Samuli Laine, Timo Aila
2017 arXiv   pre-print
In this paper, we present a simple and efficient method for training deep neural networks in a semi-supervised setting where only a small portion of training data is labeled.  ...  Using our method, we set new records for two standard semi-supervised learning benchmarks, reducing the (non-augmented) classification error rate from 18.44% to 7.05% in SVHN with 500 labels and from 18.63%  ...  The ramp-down curve was similar to the ramp-up curve but time-reversed and with a scaling constant of 12.5 instead of 5. All networks were trained for 300 epochs with minibatch size of 100.  ... 
arXiv:1610.02242v3 fatcat:x4urhkedibd7za6yqlp2v7hj2a

Liver and Tumor Segmentation Techniques for CT Abdominal Images

2019 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
Some of the techniques to segment liver from CT scanned abdominal image and to segment tumor from the liver are discussed.  ...  Image segmentation is one of the important step in digital image processing where the images are partitioned into different segments based on several properties like brightness, contrast, intensity and  ...  a) Parametric Deformable Models: The other name for Parametric Deformable Models (PDM) is called as snakes. Snakes are the curves which lie within the domain of an image.  ... 
doi:10.35940/ijitee.b1051.1292s19 fatcat:6a2zt2b2vjf27nz6hvxofk6ssy

Learning with Inadequate and Incorrect Supervision [article]

Chen Gong and Hengmin Zhang and Jian Yang and Dacheng Tao
2019 arXiv   pre-print
state-of-the-art methods in the presence of label noise and label scarcity.  ...  To address label inaccuracy, Graph Trend Filtering (GTF) and Smooth Eigenbase Pursuit (SEP) are adopted to filter out the initial noisy labels.  ...  Note that the labels of two labeled examples are incorrect and they form the noisy labels. (b)∼(e) present the classification results of different settings.  ... 
arXiv:1902.07429v1 fatcat:526u2mxm2bg2nd3nlhd3vbosyq

Fish-Eye Camera Video Processing and Trajectory Estimation Using 3D Human Models [chapter]

Konstantina Kottari, Kostas Delibasis, Vassilis Plagianakos, Ilias Maglogiannis
2014 IFIP Advances in Information and Communication Technology  
authors would like to thank the European Union (European Social Fund ESF) and Greek national funds for financially supporting this work through the Operational Program "Education and Lifelong Learning" of  ...  The parametrical equation of GC (4) can be simplified by using a piecewise straight line as curve C 1 , each segment of which is defined by vector (a 0 , b 0 , c 0 ) and by assigning these ellipses as  ...  The same process is repeated, along hands and legs -see Each intersection is estimated for approximating an ellipse with its semi-axes semi a , semi b parallel to X and Y axis of coordinate system, as  ... 
doi:10.1007/978-3-662-44654-6_38 fatcat:gcrl3miasffrvldvcylnpwxarq

Semi-parametric estimation of shifts

Fabrice Gamboa, Jean-Michel Loubes, Elie Maza
2007 Electronic Journal of Statistics  
Fourier transform enables to transform this statistical problem into a semi-parametric framework. We study the convergence of the estimator and provide its asymptotic behavior.  ...  Moreover, we use the method in the applied case of velocity curve forecasting.  ...  . / Semi-parametric estimation of shifts  ... 
doi:10.1214/07-ejs026 fatcat:t7aiwdmyc5gnfckttw3tra5pnq

Semi-Supervised Active Learning for Sound Classification in Hybrid Learning Environments

Wenjing Han, Eduardo Coutinho, Huabin Ruan, Haifeng Li, Björn Schuller, Xiaojie Yu, Xuan Zhu, Friedhelm Schwenker
2016 PLoS ONE  
Coping with scarcity of labeled data is a common problem in sound classification tasks.  ...  A reduction of 52.2% in human labeled instances is achieved in both of the pool-based and stream-based scenarios on a sound classification task considering 16,930 sound instances.  ...  Acknowledgments The authors would like to thank Zixing Zhang for his help with collecting data, and Jun Deng for their valuable feedback on the study design and on earlier versions of this manuscript.  ... 
doi:10.1371/journal.pone.0162075 pmid:27627768 pmcid:PMC5023122 fatcat:a3cerzf6s5exlbfktgqore7xue
« Previous Showing results 1 — 15 out of 6,783 results