Filters








20,762 Hits in 5.2 sec

Fast Cell Segmentation Using Scalable Sparse Manifold Learning and Affine Transform-Approximated Active Contour [chapter]

Fuyong Xing, Lin Yang
2015 Lecture Notes in Computer Science  
Efficient and effective cell segmentation of neuroendocrine tumor (NET) in whole slide scanned images is a difficult task due to a large number of cells.  ...  Followed by a shape clustering on the manifold, a novel affine transform-approximated active contour model is derived to deform contours without solving a large amount of computationally-expensive Euler-Lagrange  ...  Introduction Effective and efficient cell segmentation of pancreatic neuroendocrine tumor (NET) is a prerequisite for quantitative image analyses such as Ki67 counting.  ... 
doi:10.1007/978-3-319-24574-4_40 pmid:27924317 pmcid:PMC5136469 fatcat:n4zj7lxe6nhhjhrxqcwwg5rcfe

Biconvex Relaxation for Semidefinite Programming in Computer Vision [article]

Sohil Shah, Abhay Kumar, Carlos Castillo, David Jacobs, Christoph Studer, Tom Goldstein
2016 arXiv   pre-print
We showcase the efficacy of our approach on three applications in computer vision, namely segmentation, co-segmentation, and manifold metric learning.  ...  We propose a general framework to approximately solve large-scale semidefinite problems (SDPs) at low complexity.  ...  and metric learning on manifolds.  ... 
arXiv:1605.09527v2 fatcat:6ndyravgwndwpcpeyvckxnoygi

Biconvex Relaxation for Semidefinite Programming in Computer Vision [chapter]

Sohil Shah, Abhay Kumar Yadav, Carlos D. Castillo, David W. Jacobs, Christoph Studer, Tom Goldstein
2016 Lecture Notes in Computer Science  
We showcase the efficacy of our approach on three applications in computer vision, namely segmentation, co-segmentation, and manifold metric learning.  ...  We propose a general framework to approximately solve large-scale semidefinite problems (SDPs) at low complexity.  ...  and metric learning on manifolds.  ... 
doi:10.1007/978-3-319-46466-4_43 fatcat:egbti5p5crc3fnqgvewfuwixai

(Hyper)-graphical models in biomedical image analysis

Nikos Paragios, Enzo Ferrante, Ben Glocker, Nikos Komodakis, Sarah Parisot, Evangelia I. Zacharaki
2016 Medical Image Analysis  
This is mostly due the development of efficient learning and inference algorithms which allow better and richer modeling of image and visual understanding tasks.  ...  a variety of problems in biomedical image analysis.  ...  an approach that is metric-free, modular, scalable and computationally efficient.  ... 
doi:10.1016/j.media.2016.06.028 pmid:27377331 fatcat:f2jy24hynva2fgjfyvrcmfs6e4

Automatic Road Segmentation from High Resolution Satellite Images using Encoder-Decoder Network

2019 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
The experiments on the Massachusetts roads dataset show that our proposed model can produce precise segmentation results than other state-of-the-art models without being computationally expensive.  ...  One path of the decode module approximates the coarse spatial features using upsampling network. The other path uses Atrous spatial pyramid pooling module to extract multi scale context information.  ...  With the advancements in the area of deep learning especially in image classification they have become the best viable option for image segmentations tasks by achieving close to real and state of the art  ... 
doi:10.35940/ijitee.j9132.0881019 fatcat:frfpil3f2nayrfjb6qnkrfwt7m

Mobile-PolypNet: Lightweight Colon Polyp Segmentation Network for Low-Resource Settings

Ranit Karmakar, Saeid Nooshabadi
2022 Journal of Imaging  
We also performed ablation testing on the model to test various aspects of the autoencoder architecture. We performed the model evaluation by using most of the common image-segmentation metrics.  ...  This paper proposes a lightweight deep learning model for colorectal polyp segmentation that achieved state-of-the-art accuracy while significantly reducing the model size and complexity.  ...  Professor and Ophthalmologist at Johns Hopkins University School of Medicine, who reviewed and provided feedback on our work. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/jimaging8060169 pmid:35735968 pmcid:PMC9225047 fatcat:y6t2yyp5p5awljxm4q7l2cbrla

Deep Learning for Efficient Reconstruction of High-Resolution Turbulent DNS Data [article]

Pranshu Pant, Amir Barati Farimani
2021 arXiv   pre-print
Our model efficiently reconstructs the high-fidelity DNS data from the LES like low-resolution solutions while yielding good reconstruction metrics.  ...  Through this paper, we introduce a novel deep learning framework SR-DNS Net, which aims to mitigate this inherent trade-off between solution fidelity and computational complexity by leveraging deep learning  ...  Pixel-Shuffle is highly efficient in up-sampling images and can decrease the computational complexity of the network logarithmically in the image dimension compared to the conventional convolutional methods  ... 
arXiv:2010.11348v2 fatcat:may6m754qrcbto3nr3jgzg732i

Left Ventricle Segmentation via Graph Cut Distribution Matching [chapter]

Ismail Ben Ayed, Kumaradevan Punithakumar, Shuo Li, Ali Islam, Jaron Chong
2009 Lecture Notes in Computer Science  
Unlike related active contour methods, it does not compute iterative updates of computationally expensive kernel densities.  ...  model learned from the first frame.  ...  approximately 300 KDEs. complex learning/modeling of geometric characteristics and the need of a training set and (2) explicit optimization with respect to geometric transformations.  ... 
doi:10.1007/978-3-642-04271-3_109 fatcat:p57es7pwvnbvxmqu23y546zq2a

MalariaNet: A Computationally Efficient Convolutional Neural Network Architecture for Automated Malaria Detection

Rohan Bhansali, Loudoun Academy of Science
2020 International Journal of Engineering Research and  
We develop a computationally efficient, relatively shallow neural network architecture that can diagnose malaria from cell images obtained from thin blood smear slides.  ...  of running elaborate models makes deep learning based detection methods inaccessible in remote areas of the world.  ...  It is the simplest method of multivariate interpolation, which also makes it the most computationally efficient.  ... 
doi:10.17577/ijertv9is120158 fatcat:vuecffgikbdpjnzusxy5pbwxve

REF-Net: Robust, Efficient, and Fast Network for Semantic Segmentation Applications Using Devices With Limited Computational Resources

Bekhzod Olimov, Jeonghong Kim, Anand Paul
2021 IEEE Access  
Thanks to emergence of Deep Learning (DL) techniques, the segmentation models enhanced their accuracy.  ...  Moreover, despite its notable efficiency in terms of memory and time, the REF-Net attained superior results in several segmentation evaluation metrics that showed roughly 2%, 4%, and 3% increase in pixel  ...  This strategy assists to reduce computational complexity by applying 1 × 1 convolution to reduce the depth of an incoming image.  ... 
doi:10.1109/access.2021.3052791 fatcat:27ui6f4ejnfhlh57y7y66gbzsa

Prostate Segmentation with Texture Enhanced Active Appearance Model

S Ghose, A Oliver, R Marti, Xavier Llado, J Freixenet, J C Vilanova, F Meriaudeau
2010 2010 Sixth International Conference on Signal-Image Technology and Internet Based Systems  
Our proposed method performs computationally efficient accurate multi-modal prostate segmentation in presence of intensity heterogeneities and imaging artifacts.  ...  Prostate contour segmented from Trans Rectal Ultra Sound (TRUS) and Magnetic Resonance (MR) images could improve inter-modality registration accuracy and reduce computational complexity of the procedure  ...  ACKNOWLEDGMENT This research is a part of the PROSCAN project of the VICOROB laboratory of University of Girona, Catalunya, Spain.  ... 
doi:10.1109/sitis.2010.14 dblp:conf/sitis/GhoseOMLFVM10 fatcat:defqmwbe4vc7rlc5a55isqff7i

ACTIVE LEARNING ON LARGE HYPERSPECTRAL DATASETS: A PREPROCESSING METHOD

R. Thoreau, V. Achard, L. Risser, B. Berthelot, X. Briottet
2022 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
Machine learning algorithms demonstrated promising results for hyperspectral semantic segmentation. However, they strongly rely on the quality of training datasets.  ...  In the machine learning community, recent active learning methods have overcome the performance of conventional algorithms but do not always scale to large remote sensing images.  ...  use of machine learning models to large and complex images.  ... 
doi:10.5194/isprs-archives-xliii-b3-2022-435-2022 fatcat:tt24h7f7rbdxdlq6ytp7otaehy

Transferability Metrics for Selecting Source Model Ensembles [article]

Andrea Agostinelli, Jasper Uijlings, Thomas Mensink, Vittorio Ferrari
2022 arXiv   pre-print
Since fine-tuning all possible ensembles is computationally prohibitive, we aim at predicting performance on the target dataset using a computationally efficient transferability metric.  ...  We propose several new transferability metrics designed for this task and evaluate them in a challenging and realistic transfer learning setup for semantic segmentation: we create a large and diverse pool  ...  The number of datapoints in semantic segmentation is approximately 6 orders of magnitude higher than in image classification.  ... 
arXiv:2111.13011v2 fatcat:gnqqhexscvcufiairhr2w5yjkq

Learning Densities in Feature Space for Reliable Segmentation of Indoor Scenes [article]

Nicolas Marchal, Charlotte Moraldo, Roland Siegwart, Hermann Blum, Cesar Cadena, Abel Gawel
2020 arXiv   pre-print
Deep learning has enabled remarkable advances in scene understanding, particularly in semantic segmentation tasks.  ...  As our method does not explicitly learn the representation of individual objects, its performance generalizes well outside of the training examples.  ...  Unfortunately, this method requires kNN lookups for every image patch, which is computationally inefficient. 2) Flow-Based Density Estimation: Flow-based approaches have proven useful for estimating complex  ... 
arXiv:1908.00448v4 fatcat:teafty5c6jd6fnotvuaa72pthm

Lightweight Vision Transformer with Cross Feature Attention [article]

Youpeng Zhao, Huadong Tang, Yingying Jiang, Yong A, Qiang Wu
2022 arXiv   pre-print
On Cityscapes dataset, with only a simple all-MLP decoder, XFormer achieves mIoU of 78.5 and FPS of 15.3, surpassing state-of-the-art lightweight segmentation networks.  ...  of parameters.  ...  Linformer [54] replaces self-attention with low-rank approximation operation, but large batch matrix decomposition is still computationally expensive and the efficiency gain is only noticeable for large  ... 
arXiv:2207.07268v1 fatcat:fpg3xoqpqjhwno3a62e4eaedpe
« Previous Showing results 1 — 15 out of 20,762 results