The Experts below are selected from a list of 111078 Experts worldwide ranked by ideXlab platform
Iasonas Kokkinos - One of the best experts on this subject based on the ideXlab platform.
-
Segmentation-Aware Convolutional Networks Using Local Attention Masks
arXiv: Computer Vision and Pattern Recognition, 2017Co-Authors: Adam W. Harley, Konstantinos G. Derpanis, Iasonas KokkinosAbstract:We introduce an approach to integrate Segmentation Information within a convolutional neural network (CNN). This counter-acts the tendency of CNNs to smooth Information across regions and increases their spatial precision. To obtain Segmentation Information, we set up a CNN to provide an embedding space where region co-membership can be estimated based on Euclidean distance. We use these embeddings to compute a local attention mask relative to every neuron position. We incorporate such masks in CNNs and replace the convolution operation with a "Segmentation-aware" variant that allows a neuron to selectively attend to inputs coming from its own region. We call the resulting network a Segmentation-aware CNN because it adapts its filters at each image point according to local Segmentation cues. We demonstrate the merit of our method on two widely different dense prediction tasks, that involve classification (semantic Segmentation) and regression (optical flow). Our results show that in semantic Segmentation we can match the performance of DenseCRFs while being faster and simpler, and in optical flow we obtain clearly sharper responses than networks that do not use local attention masks. In both cases, Segmentation-aware convolution yields systematic improvements over strong baselines. Source code for this work is available online at this http URL.
-
ICCV - Segmentation-Aware Convolutional Networks Using Local Attention Masks
2017 IEEE International Conference on Computer Vision (ICCV), 2017Co-Authors: Adam W. Harley, Konstantinos G. Derpanis, Iasonas KokkinosAbstract:We introduce an approach to integrate Segmentation Information within a convolutional neural network (CNN). This counter-acts the tendency of CNNs to smooth Information across regions and increases their spatial precision. To obtain Segmentation Information, we set up a CNN to provide an embedding space where region co-membership can be estimated based on Euclidean distance. We use these embeddings to compute a local attention mask relative to every neuron position. We incorporate such masks in CNNs and replace the convolution operation with a “Segmentation-aware” variant that allows a neuron to selectively attend to inputs coming from its own region. We call the resulting network a Segmentation-aware CNN because it adapts its filters at each image point according to local Segmentation cues, while at the same time remaining fully-convolutional. We demonstrate the merit of our method on two widely different dense prediction tasks, that involve classification (semantic Segmentation) and regression (optical flow). Our results show that in semantic Segmentation we can replace DenseCRF inference with a cascade of Segmentation-aware filters, and in optical flow we obtain clearly sharper responses than the ones obtained with comparable networks that do not use Segmentation. In both cases Segmentation-aware convolution yields systematic improvements over strong baselines.
Changshui Zhang - One of the best experts on this subject based on the ideXlab platform.
-
Deep transformer: A framework for 2D text image rectification from planar transformations
Neurocomputing, 2018Co-Authors: Chengzhe Yan, Changshui ZhangAbstract:Abstract In this paper, a novel neural network architecture is proposed to rectify text images with mild assumptions. A new dataset of text images is collected to verify our model. We explored the capability of deep neural network in learning geometric transformation and found the model are sensitive to the text image without explicit supervised Segmentation Information. Experiments show the architecture proposed can restore planar transformations with wonderful robustness and effectiveness.
-
A DNN Framework For Text Image Rectification From Planar Transformations.
arXiv: Computer Vision and Pattern Recognition, 2016Co-Authors: Chengzhe Yan, Changshui ZhangAbstract:In this paper, a novel neural network architecture is proposed attempting to rectify text images with mild assumptions. A new dataset of text images is collected to verify our model and open to public. We explored the capability of deep neural network in learning geometric transformation and found the model could segment the text image without explicit supervised Segmentation Information. Experiments show the architecture proposed can restore planar transformations with wonderful robustness and effectiveness.
Adam W. Harley - One of the best experts on this subject based on the ideXlab platform.
-
Segmentation-Aware Convolutional Networks Using Local Attention Masks
arXiv: Computer Vision and Pattern Recognition, 2017Co-Authors: Adam W. Harley, Konstantinos G. Derpanis, Iasonas KokkinosAbstract:We introduce an approach to integrate Segmentation Information within a convolutional neural network (CNN). This counter-acts the tendency of CNNs to smooth Information across regions and increases their spatial precision. To obtain Segmentation Information, we set up a CNN to provide an embedding space where region co-membership can be estimated based on Euclidean distance. We use these embeddings to compute a local attention mask relative to every neuron position. We incorporate such masks in CNNs and replace the convolution operation with a "Segmentation-aware" variant that allows a neuron to selectively attend to inputs coming from its own region. We call the resulting network a Segmentation-aware CNN because it adapts its filters at each image point according to local Segmentation cues. We demonstrate the merit of our method on two widely different dense prediction tasks, that involve classification (semantic Segmentation) and regression (optical flow). Our results show that in semantic Segmentation we can match the performance of DenseCRFs while being faster and simpler, and in optical flow we obtain clearly sharper responses than networks that do not use local attention masks. In both cases, Segmentation-aware convolution yields systematic improvements over strong baselines. Source code for this work is available online at this http URL.
-
ICCV - Segmentation-Aware Convolutional Networks Using Local Attention Masks
2017 IEEE International Conference on Computer Vision (ICCV), 2017Co-Authors: Adam W. Harley, Konstantinos G. Derpanis, Iasonas KokkinosAbstract:We introduce an approach to integrate Segmentation Information within a convolutional neural network (CNN). This counter-acts the tendency of CNNs to smooth Information across regions and increases their spatial precision. To obtain Segmentation Information, we set up a CNN to provide an embedding space where region co-membership can be estimated based on Euclidean distance. We use these embeddings to compute a local attention mask relative to every neuron position. We incorporate such masks in CNNs and replace the convolution operation with a “Segmentation-aware” variant that allows a neuron to selectively attend to inputs coming from its own region. We call the resulting network a Segmentation-aware CNN because it adapts its filters at each image point according to local Segmentation cues, while at the same time remaining fully-convolutional. We demonstrate the merit of our method on two widely different dense prediction tasks, that involve classification (semantic Segmentation) and regression (optical flow). Our results show that in semantic Segmentation we can replace DenseCRF inference with a cascade of Segmentation-aware filters, and in optical flow we obtain clearly sharper responses than the ones obtained with comparable networks that do not use Segmentation. In both cases Segmentation-aware convolution yields systematic improvements over strong baselines.
Dwarikanath Mahapatra - One of the best experts on this subject based on the ideXlab platform.
-
DART/DCL@MICCAI - Registration of Histopathology Images Using Self Supervised Fine Grained Feature Maps.
Domain Adaptation and Representation Transfer and Distributed and Collaborative Learning, 2020Co-Authors: James Tong, Dwarikanath Mahapatra, C. Paul Bonnington, Tom DrummondAbstract:Image registration is an important part of many clinical workflows and inclusion of Segmentation Information of structures of interest improves registration performance. We propose to integrate Segmentation Information in a registration framework using fine grained feature maps obtained in a self supervised manner. Self supervised feature maps enables use of Segmentation Information despite the unavailability of manual Segmentations. Experimental results show our approach effectively replaces manual Segmentation maps and demonstrate the possibility of obtaining state of the art registration performance in real world cases where manual Segmentation maps are unavailable.
-
Registration of Histopathogy Images Using Structural Information From Fine Grained Feature Maps.
arXiv: Image and Video Processing, 2020Co-Authors: Dwarikanath MahapatraAbstract:Registration is an important part of many clinical workflows and factually, including Information of structures of interest improves registration performance. We propose a novel approach of combining Segmentation Information in a registration framework using self supervised Segmentation feature maps extracted using a pre-trained Segmentation network followed by clustering. Using self supervised feature maps enables us to use Segmentation Information despite the unavailability of manual Segmentations. Experimental results show our approach effectively replaces manual Segmentation maps and demonstrate the possibility of obtaining state of the art registration performance in real world cases where manual Segmentation maps are unavailable.
-
Training Data Independent Image Registration With GANs Using Transfer Learning And Segmentation Information
arXiv: Computer Vision and Pattern Recognition, 2019Co-Authors: Dwarikanath MahapatraAbstract:Registration is an important task in automated medical image analysis. Although deep learning (DL) based image registration methods out perform time consuming conventional approaches, they are heavily dependent on training data and do not generalize well for new images types. We present a DL based approach that can register an image pair which is different from the training images. This is achieved by training generative adversarial networks (GANs) in combination with Segmentation Information and transfer learning. Experiments on chest Xray and brain MR images show that our method gives better registration performance over conventional methods.
-
ISBI - Training Data Independent Image Registration with Gans Using Transfer Learning and Segmentation Information
2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019), 2019Co-Authors: Dwarikanath MahapatraAbstract:Registration is an important task in automated medical image analysis. Although deep learning (DL) based image registration methods out perform time consuming conventional approaches, they are heavily dependent on training data and do not generalize well for new images types. We present a DL based approach that can register an image pair which is different from the training images. This is achieved by training generative adversarial networks (GANs) in combination with Segmentation Information and transfer learning. Experiments on chest Xray and brain MR images show that our method gives better registration performance over conventional methods.
-
Medical Imaging: Image Processing - Groupwise registration of dynamic cardiac perfusion images using temporal dynamics and Segmentation Information
Medical Imaging 2012: Image Processing, 2012Co-Authors: Dwarikanath MahapatraAbstract:ABSTRACT We propose a groupwise registration method for cardiac perfusion images using temporal Information. Perfusionimage characteristics make groupwise registration especially attractive, and registration accuracy is improvedby including Segmentation Information. Registration aims to maximize the smoothness of the intensity signalwhile Segmentation minimizes a pixels intensity dierence with other pixels having the same Segmentation label.We use B-splines to optimize the cost function that includes registration and Segmentation Information. Themethod was tested on real patient datasets. Due to the use of temporal Information for groupwise registrationour method shows lower registration error and higher Segmentation accuracy than conventional methods whichperform registration and Segmentation separately.Keywords: Groupwise registration, Segmentation, temporal Information, cardiac, perfusion, MRI 1. INTRODUCTION Dynamic contrast enhanced (DCE) magnetic resonance (MR) images (or perfusion MRI) has developed as apopular non-invasive tool for the functional analysis of internal organs. Contrast agent is injected intravenouslyinto the patient and a series of MR images are acquired over a period of time. As the contrast agent owsthrough the blood stream, the intensity of corresponding regions increases. Since the image acquisition processcan take up to 20 minutes, patient movement is inevitable. Additionally, there is elastic deformation of cardiactissues due to patient breathing which have to be corrected. The perfusion images are characterized by rapidintensity change over time, low spatial resolution and noise, posing challenges to registration. Previous techniquesmostly employed a pairwise registration approach, i.e., all images of a sequence are individually registered to axed reference image.
Chengzhe Yan - One of the best experts on this subject based on the ideXlab platform.
-
Deep transformer: A framework for 2D text image rectification from planar transformations
Neurocomputing, 2018Co-Authors: Chengzhe Yan, Changshui ZhangAbstract:Abstract In this paper, a novel neural network architecture is proposed to rectify text images with mild assumptions. A new dataset of text images is collected to verify our model. We explored the capability of deep neural network in learning geometric transformation and found the model are sensitive to the text image without explicit supervised Segmentation Information. Experiments show the architecture proposed can restore planar transformations with wonderful robustness and effectiveness.
-
A DNN Framework For Text Image Rectification From Planar Transformations.
arXiv: Computer Vision and Pattern Recognition, 2016Co-Authors: Chengzhe Yan, Changshui ZhangAbstract:In this paper, a novel neural network architecture is proposed attempting to rectify text images with mild assumptions. A new dataset of text images is collected to verify our model and open to public. We explored the capability of deep neural network in learning geometric transformation and found the model could segment the text image without explicit supervised Segmentation Information. Experiments show the architecture proposed can restore planar transformations with wonderful robustness and effectiveness.