Filters








52,059 Hits in 4.3 sec

Learning to Reconstruct 3D Human Pose and Shape via Model-fitting in the Loop [article]

Nikos Kolotouros, Georgios Pavlakos, Michael J. Black, Kostas Daniilidis
2019 arXiv   pre-print
The deep network initializes an iterative optimization routine that fits the body model to 2D joints within the training loop, and the fitted estimate is subsequently used to supervise the network.  ...  of supervision.  ...  While MJB is a part-time employee of Amazon, his research was performed solely at, and funded solely by, MPI. MJB has financial interests in Amazon and Meshcapade GmbH.  ... 
arXiv:1909.12828v1 fatcat:z3qs56umanfapaszx2gunuvra4

Unsupervised Feature Learning via Non-parametric Instance Discrimination

Zhirong Wu, Yuanjun Xiong, Stella X. Yu, Dahua Lin
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
We formulate this intuition as a non-parametric classification problem at the instance-level, and use noisecontrastive estimation to tackle the computational challenges imposed by the large number of instance  ...  By fine-tuning the learned feature, we further obtain competitive results for semi-supervised learning and object detection tasks.  ...  Noise-Contrastive Estimation Computing the non-parametric softmax in Eq. (2) is cost prohibitive when the number of classes n is very large, e.g. at the scale of millions.  ... 
doi:10.1109/cvpr.2018.00393 dblp:conf/cvpr/WuXYL18 fatcat:gz7a6t6yxbhktavshli6o32br4

Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination [article]

Zhirong Wu, Yuanjun Xiong, Stella Yu, Dahua Lin
2018 arXiv   pre-print
We formulate this intuition as a non-parametric classification problem at the instance-level, and use noise-contrastive estimation to tackle the computational challenges imposed by the large number of  ...  By fine-tuning the learned feature, we further obtain competitive results for semi-supervised learning and object detection tasks.  ...  Noise-Contrastive Estimation Computing the non-parametric softmax in Eq. (2) is cost prohibitive when the number of classes n is very large, e.g. at the scale of millions.  ... 
arXiv:1805.01978v1 fatcat:3ilhvltoinfo5ckxynd5fhjtdq

Page 94 of Journal of Research and Practice in Information Technology Vol. 26, Issue 3 [page]

1994 Journal of Research and Practice in Information Technology  
Supervised methods require a human operator to extract samples of each tissue type from the images in order to initially estimate the cluster locations.  ...  In supervised methods, the parameters of the model are estimated from training data of which the true classification (tissue type) is known.  ... 

PhyCOM: A Multi-Layer Parametric Network for Joint Linear Impairments Compensation and Symbol Detection [article]

Vincent Choqueuse, Alexandru Frunza, Stéphane Azou, Pascal Morel
2022 arXiv   pre-print
This model is composed of widely linear parametric layers that describe the input-output relationship of the front-end impairments and channel effects.  ...  Because of the small number of network parameters, a PhyCOM network can be trained efficiently using sophisticated optimization algorithms and a limited number of pilot symbols.  ...  Supervised Training Let us denote the initial network parameters by θ = [θ T 1 , • • • , θ T L ] T .  ... 
arXiv:2203.00266v1 fatcat:qfp6lo5qjrhm5ony3ux7pabw7q

TexturePose: Supervising Human Mesh Estimation with Texture Consistency [article]

Georgios Pavlakos, Nikos Kolotouros, Kostas Daniilidis
2019 arXiv   pre-print
This seemingly insignificant and often overlooked cue goes a long way for model-based pose estimation. The parametric model we employ allows us to compute a texture map for each frame.  ...  This work addresses the problem of model-based human pose estimation.  ...  The initial baseline (first row) is the same as in Table 1 , and uses full 3D ground truth from Human3.6M for training.  ... 
arXiv:1910.11322v1 fatcat:d6i2llmnzbfxplv4j7nym3ll54

Bayesian Example Based Segmentation using a Hybrid Energy Model

Claire Gallagher, Anil Kokaram
2007 2007 IEEE International Conference on Image Processing  
This paper describes a supervised segmentation algorithm which draws inspiration from recent advances in non-parametric texture synthesis.  ...  The suitability of the wavelet transform for texture modelling is highlighted and an outlier class condition is introduced as a means to increase the flexibility of the algorithm.  ...  This problem affects all supervised algorithms in that the training set must be rich enough for segmentation to be useful.  ... 
doi:10.1109/icip.2007.4379087 dblp:conf/icip/GallagherK07 fatcat:w6hop47tdffkvad57ntlmbtww4

Combining parametric and non-parametric algorithms for a partially unsupervised classification of multitemporal remote-sensing images

Lorenzo Bruzzone, Roberto Cossu, Gianni Vernazza
2002 Information Fusion  
The proposed system is composed of an ensemble of classifiers that, once trained in a supervised way on a specific image of a given area, can be retrained in an unsupervised way to classify a new image  ...  In this context, two techniques are presented for the unsupervised updating of the parameters of a maximum-likelihood classifier and a radial basis function neural-network classifier, on the basis of the  ...  All the components of # n 2 are initialized according to the values obtained in a supervised way on the t 1 image.  ... 
doi:10.1016/s1566-2535(02)00091-x fatcat:73rdzxak2zbe5hho73u2darrai

Segmentation of Brain Tumors in Multi-parametric MR Images via Robust Statistic Information Propagation [chapter]

Hongming Li, Ming Song, Yong Fan
2011 Lecture Notes in Computer Science  
The proposed method has been applied to 3D multi-parametric MR images with tumors of different sizes and locations.  ...  weights of graph edges.  ...  This study was supported in part by the National Science Foundation of China (Grant number: 30970770) and the Hundred Talents Programs, Chinese Academy of Sciences.  ... 
doi:10.1007/978-3-642-19282-1_48 fatcat:kcetrkcikjatpj6cgkdjxrjakq

Convolutional Mesh Regression for Single-Image Human Shape Reconstruction

Nikos Kolotouros, Georgios Pavlakos, Kostas Daniilidis
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
This paper addresses the problem of 3D human pose and shape estimation from a single image.  ...  Previous approaches consider a parametric model of the human body, SMPL, and attempt to regress the model parameters that give rise to a mesh consistent with image evidence.  ...  Now, we train the network using two forms of supervision.  ... 
doi:10.1109/cvpr.2019.00463 dblp:conf/cvpr/KolotourosPD19 fatcat:exyr5zm7dvdbrdge6vbeqzehme

Convolutional Mesh Regression for Single-Image Human Shape Reconstruction [article]

Nikos Kolotouros, Georgios Pavlakos, Kostas Daniilidis
2019 arXiv   pre-print
This paper addresses the problem of 3D human pose and shape estimation from a single image.  ...  Previous approaches consider a parametric model of the human body, SMPL, and attempt to regress the model parameters that give rise to a mesh consistent with image evidence.  ...  Now, we train the network using two forms of supervision.  ... 
arXiv:1905.03244v1 fatcat:cf7wrt4if5gd5c4rv53uon5zpq

Robust segmentation using non-parametric snakes with multiple cues for applications in radiation oncology

Jayashree Kalpathy-Cramer, Umut Ozertem, William Hersh, Martin Fuss, Deniz Erdogmus, Josien P. W. Pluim, Benoit M. Dawant
2009 Medical Imaging 2009: Image Processing  
It allows multiple cues to be incorporated easily for the purposes of estimating the edge probability density.  ...  We compare the results using a non-parametric snake technique with a gold standard consisting of manually delineated structures for tumors as well as normal organs.  ...  Apiradee Srisuthep, Research Fellow in the Department of Radiation Medicine, OHSU, Portland, OR for her help in acquiring the manually labeled data.  ... 
doi:10.1117/12.812712 dblp:conf/miip/Kalpathy-Cramer09 fatcat:tilkihcudfax3ey7uld7fd63sa

PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop [article]

Hongwen Zhang, Yating Tian, Xinchi Zhou, Wanli Ouyang, Yebin Liu, Limin Wang, Zhenan Sun
2021 arXiv   pre-print
To reduce noise and enhance the reliability of these evidences, an auxiliary pixel-wise supervision is imposed on the feature encoder, which provides mesh-image correspondence guidance for our network  ...  of the reconstruction.  ...  Human3.6M [17] is commonly used as the benchmark dataset for 3D human pose estimation, consisting of 3.6 million video frames captured in the controlled environment.  ... 
arXiv:2103.16507v4 fatcat:mlew2hk7dbhvpexr7t76rjxe4a

Semi-Supervised Image Classification in Likelihood Space

Rong Duan, Wei Jiang, Hong Man
2006 2006 International Conference on Image Processing  
Semi-supervised learning : using large amount of unlabeled training data to help limited amount of labeled training data to improve classification performance. parametric generative mixture models approach  ...  [5], simulation Conclusion: When model mis-specified , unlabeled data help to improve classification performance only when the estimation error for labeled training data is bigger than model error and  ... 
doi:10.1109/icip.2006.312634 dblp:conf/icip/DuanJM06 fatcat:mbycjxft25aclncputmhlwpwte

Learning Unsupervised Word Mapping by Maximizing Mean Discrepancy [article]

Pengcheng Yang, Fuli Luo, Shuangzhi Wu, Jingjing Xu, Dongdong Zhang, Xu Sun
2018 arXiv   pre-print
or intermediate density estimation.  ...  Most previous work of this task adopts parametric metrics to measure distribution differences, which typically requires a sophisticated alternate optimization process, either in the form of minmax game  ...  Compression and Initialization At the training stage, Eq. (3) the MMD. A reliable estimation of Eq. (3) generally requires the size of the mini-batch to be proportional to the dimension.  ... 
arXiv:1811.00275v1 fatcat:ehnzu3z5abh5va5ogoyvyjqe2e
« Previous Showing results 1 — 15 out of 52,059 results