Filters








4,570 Hits in 4.0 sec

Learning Conditional Deformable Templates with Convolutional Networks [article]

Adrian V. Dalca, Marianne Rakic, John Guttag, Mert R. Sabuncu
2019 arXiv   pre-print
In this work, we present a probabilistic model and efficient learning strategy that yields either universal or conditional templates, jointly with a neural network that provides efficient alignment of  ...  We develop a learning framework for building deformable templates, which play a fundamental role in many image analysis and computational anatomy tasks.  ...  We model the conditional template network architecture as a decoder with a dense layer followed by several upsampling and convolutional levels.  ... 
arXiv:1908.02738v2 fatcat:r3b4yiyghzbipftp6csb6orbea

Whole Heart Mesh Generation For Image-Based Computational Simulations By Learning Free-From Deformations [article]

Fanwei Kong, Shawn C. Shadden
2021 arXiv   pre-print
Our approach learns to deform a template mesh to the input image data by predicting displacements of multi-resolution control point grids.  ...  We propose a novel deep learning approach to reconstruct simulation-ready whole heart meshes from volumetric image data.  ...  In contrast to learning deformation on a template mesh, [22] proposed to learn the space deformation by predicting the displacements of a control point grid to deform template meshes of the lung.  ... 
arXiv:2107.10839v1 fatcat:dt2r5bd5w5hilp5wznnlapixym

Generative Adversarial Registration for Improved Conditional Deformable Templates [article]

Neel Dey, Mengwei Ren, Adrian V. Dalca, Guido Gerig
2022 arXiv   pre-print
We reformulate deformable registration and conditional template estimation as an adversarial game wherein we encourage realism in the moved templates with a generative adversarial registration framework  ...  Current conventional and deep network-based methods for template construction use only regularized registration objectives and often yield templates with blurry and/or anatomically implausible appearance  ...  A template generation network (a) processes an array of learned parameters with a convolutional decoder whose feature-wise affine parameters are learned from input conditions to generate a conditional  ... 
arXiv:2105.04349v2 fatcat:lg32vthf6nchlaom3igzpfnvmi

Deep Learning Improves Template Matching by Normalized Cross Correlation [article]

Davit Buniatyan, Thomas Macrina, Dodam Ih, Jonathan Zung, H. Sebastian Seung
2017 arXiv   pre-print
We improve the robustness of this algorithm by preprocessing images with "siamese" convolutional networks trained to maximize the contrast between NCC values of true and false matches.  ...  Relative to a parameter-tuned bandpass filter, siamese convolutional networks significantly reduce false matches.  ...  We would like to explore how well this technique generalizes from one EM dataset to another, as well as to investigate if there is transfer learning that could benefit the segmentation convolutional network  ... 
arXiv:1705.08593v1 fatcat:lqnai5pbfvhldi6tmtu4lt3l4u

Deforming Autoencoders: Unsupervised Disentangling of Shape and Appearance [article]

Zhixin Shu, Mihir Sahasrabudhe, Alp Guler, Dimitris Samaras, Nikos Paragios, Iasonas Kokkinos
2018 arXiv   pre-print
As in the deformable template paradigm, shape is represented as a deformation between a canonical coordinate system ('template') and an observed image, while appearance is modeled in 'canonical', template  ...  We show experiments with expression morphing in humans, hands, and digits, face manipulation, such as shape and appearance interpolation, as well as unsupervised landmark localization.  ...  Instead of building a generic deformation model, we compose a global, affine deformation field, with a non-rigid field that is synthesized as a convolutional decoder network.  ... 
arXiv:1806.06503v1 fatcat:2y3w7ofn6fhzrkac27gsabrg74

Image-to-Graph Convolutional Network for Deformable Shape Reconstruction from a Single Projection Image [article]

M. Nakao, F. Tong, M. Nakamura, T. Matsuda
2021 arXiv   pre-print
In this paper, we propose an image-to-graph convolutional network (IGCN) for deformable shape reconstruction from a single-viewpoint projection image.  ...  The IGCN learns relationship between shape/deformation variability and the deep image features based on a deformation mapping scheme.  ...  Image-to-Graph Convolutional Network Fig . 2 shows the IGCN, which consists of a CNN that extracts perceptual features from the input image and a graph convolutional network (GCN) that learns mesh deformation  ... 
arXiv:2108.12533v2 fatcat:2yqauy47tbbkzkfk4wlwnmgije

Siamese Tracking with Adaptive Template-Updating Strategy

Zheng Xu, Haibo Luo, Bin Hui, Zheng Chang, Moran Ju
2019 Applied Sciences  
Experiment results on the DARPA dataset prove that our new tracking algorithm with the template-updating strategy prominently improved tracking accuracy regarding the deformation condition.  ...  Recently, we combined a contour-detection network and a fully convolutional Siamese tracking network to initialize a new start-up of vehicle tracking by clicking on the target, generating a contour proposal  ...  Figure 8 shows outputs from our template-updating network regarding the deformation and occlusion conditions. adaptive templates is able to deal with partial-occlusion conditions because there is less  ... 
doi:10.3390/app9183725 fatcat:teibdf7cdvbf7nf75ekhcbon3i

DeformNet: Free-Form Deformation Network for 3D Shape Reconstruction from a Single Image [article]

Andrey Kurenkov, Jingwei Ji, Animesh Garg, Viraj Mehta, JunYoung Gwak, Christopher Choy, Silvio Savarese
2017 arXiv   pre-print
DeformNet takes an image input, searches the nearest shape template from a database, and deforms the template to match the query image.  ...  We introduce a new differentiable layer for 3D data deformation and use it in DeformNet to learn a model for 3D reconstruction-through-deformation.  ...  Since it is defined on a 3D grid, FFD fits with 3D convolutional neural network and has been used for generating 3D deformation using a neural network before [38] .  ... 
arXiv:1708.04672v1 fatcat:tdbjw6rn5bch3kuidapivqgsd4

View Generalization for Single Image Textured 3D Models [article]

Anand Bhattad, Aysegul Dundar, Guilin Liu, Andrew Tao, Bryan Catanzaro
2021 arXiv   pre-print
As for generalization problems in machine learning, the difficulty is balancing single-view accuracy (cf. training error; bias) with novel view accuracy (cf. test error; variance).  ...  The network learns w, the template weight, to choose one from n templates it needs to start with for an image.  ...  Previous methods [16, 4] start with mean templates and learn to deform it. We instead use multiple templates provided by PASCAL3D+ dataset [42] , and a mean template for CUB dataset [41] .  ... 
arXiv:2106.06533v1 fatcat:z5pb3zoefbaehmpepnvwmxlw4m

Toward automatic phenotyping of developing embryos from videos

Feng Ning, D. Delhomme, Y. LeCun, F. Piano, L. Bottou, P.E. Barbano
2005 IEEE Transactions on Image Processing  
which cleans up the output of the convolutional network by learning local consistency constraints that must be satisfied by label images; (3) A set of elastic models of the embryo at various stages of  ...  The system contains three modules (1) a convolutional network trained to classify each pixel into five categories: cell wall, cytoplasm, nucleus membrane, nucleus, outside medium; (2) an Energy-Based Model  ...  Fitting dense deformable templates with Colored SOMs A second method was tested using templates with "dense" nodes.  ... 
doi:10.1109/tip.2005.852470 pmid:16190471 fatcat:grpayuf3t5datnwbuuldgbdb3a

DenseReg: Fully Convolutional Dense Shape Regression In-the-Wild [article]

Rıza Alp Güler, George Trigeorgis, Epameinondas Antonakos, Patrick Snape, Stefanos Zafeiriou, Iasonas Kokkinos
2017 arXiv   pre-print
In this paper we propose to learn a mapping from image pixels into a dense template grid through a fully convolutional network.  ...  Our system, called DenseReg, allows us to estimate dense image-to-template correspondences in a fully convolutional manner.  ...  FCNN Template Shape Deformation Free Space 1 Figure 1 : We introduce a fully convolutional neural network that regresses from the image to a "canonical", deformationfree parameterization of the face  ... 
arXiv:1612.01202v2 fatcat:fsfiinvf3bcohmmlq45ybyvboe

Synthesis of 3D MRI Brain Images With Shape and Texture Generative Adversarial Deep Neural Networks

Chee Keong Chong, Eric Tatt Wei Ho
2021 IEEE Access  
Novel images are synthesized from the generator network alone. (Top) The shape network learns a 3D deformation field from a template image.  ...  Learning both a deformation field from the MNI template and a textural mapping from the deformed template simplifies the problem to a lower-dimensional learning task; this helps high-dimensional GAN converge  ... 
doi:10.1109/access.2021.3075608 fatcat:wkmezhv2ujdsddjv4zwabxx3j4

Deep Group-wise Registration for Multi-spectral Images from Fundus Images

Tongtong Che, Yuanjie Zheng, Jinyu Cong, Yi Niu, Yanyun Jiang, Wanzhen Jiao, Bojun Zhao, Yanhui Ding
2019 IEEE Access  
The framework contains three parts: a template construction based on principal component analysis, a deformation field calculation, and a spatial transformation.  ...  INDEX TERMS Multi-spectral images, group-wise registration, deep learning, mono/multi-modal images.  ...  FIGURE 4 . 4 Proposed convolutional architecture of the group-wise registration network. Moving images are sequentially paired with the template image of size 512 × 512 as an input to the network.  ... 
doi:10.1109/access.2019.2901580 fatcat:wcs7sltwofe73oddj7elaoyavi

Deforming Autoencoders: Unsupervised Disentangling of Shape and Appearance [chapter]

Zhixin Shu, Mihir Sahasrabudhe, Rıza Alp Güler, Dimitris Samaras, Nikos Paragios, Iasonas Kokkinos
2018 Lecture Notes in Computer Science  
As in the deformable template paradigm, shape is represented as a deformation between a canonical coordinate system ('template') and an observed image, while appearance is modeled in deformation-invariant  ...  , template coordinates.  ...  Instead of building a generic deformation model, we compose a global, affine deformation field, with a non-rigid field that is synthesized as a convolutional decoder network.  ... 
doi:10.1007/978-3-030-01249-6_40 fatcat:k4c62he67bgybail7in6w7x35e

IGCN: Image-to-graph Convolutional Network for 2D/3D Deformable Registration [article]

Megumi Nakao, Mitsuhiro Nakamura, Tetsuya Matsuda
2021 arXiv   pre-print
We propose an image-to-graph convolutional network that achieves deformable registration of a 3D organ mesh for a single-viewpoint 2D projection image.  ...  The experimental results show shape prediction considering relationships among multiple organs can be used to predict respiratory motion and deformation from digitally reconstructed radiographs with clinically  ...  Thus, recent studies have investigated 3D displacement field learning using a convolutional neural network (CNN) [13] - [18] .  ... 
arXiv:2111.00484v1 fatcat:v3dxmg4dgzafbfxviq2pxfvop4
« Previous Showing results 1 — 15 out of 4,570 results