We present a novel cross-view transformer method to transfer information between unregistered views at the level of spatial feature maps.
The heavy tail and skewness characteristics of such distributions in the spectral domain allow to capture long-range covariance of the signal in the time domain.
Unsupervised domain adaptation (UDA) amounts to assigning class labels to the unlabeled instances of a dataset from a target domain, using labeled instances of a dataset from a related source domain.
We analyze two transfer learning settings: 1) unsupervised transfer, where Hologic data with soft lesion annotation at pixel level and Siemens unlabelled data are used to annotate images in the latter data; 2) weak supervised transfer, where exam level labels for images from the Siemens mammograph are available.
Multi-output Gaussian processes (MOGPs) are recently extended by using spectral mixture kernel, which enables expressively pattern extrapolation with a strong interpretation.
Multi-Task Gaussian processes (MTGPs) have shown a significant progress both in expressiveness and interpretation of the relatedness between different tasks: from linear combinations of independent single-output Gaussian processes (GPs), through the direct modeling of the cross-covariances such as spectral mixture kernels with phase shift, to the design of multivariate covariance functions based on spectral mixture kernels which model delays among tasks in addition to phase differences, and which provide a parametric interpretation of the relatedness across tasks.
By verifying the presence of dependencies between function components using Gaussian conditionals and posterior covariance, we first propose a new SM kernel variant with a time and phase delay dependency structure (SMD) and then provide a structure adaptation (SA) algorithm for the SMD.
Under the assumption that features from pre-trained deep neural networks are transferable across related domains, domain adaptation reduces to aligning source and target domain at class prediction uncertainty level.
Despite their success, state of the art methods based on this approach are either involved or unable to directly scale to data with many features.
2) How is the performance of hyperspectral image classification methods affected when using disjoint train and test sets?
Our results show that: 1) feature learning outperforms handcrafted features; 2) parameter transfer learning is beneficial in longitudinal settings; 3) using LSTM to learn the temporal dynamic of signals enhances the detection rate especially for skewed training data; 4) an ensemble of LSTMs provides more accurate and stable detectors.
Unsupervised Domain Adaptation (DA) is used to automatize the task of labeling data: an unlabeled dataset (target) is annotated using a labeled dataset (source) from a related domain.
no code implementations • 25 Feb 2017 • Mohsen Ghafoorian, Alireza Mehrtash, Tina Kapur, Nico Karssemeijer, Elena Marchiori, Mehran Pesteie, Charles R. G. Guttmann, Frank-Erik de Leeuw, Clare M. Tempany, Bram van Ginneken, Andriy Fedorov, Purang Abolmaesumi, Bram Platel, William M. Wells III
In this study, we aim to answer the following central questions regarding domain adaptation in medical image analysis: Given a fitted legacy model, 1) How much data from the new domain is required for a decent adaptation of the original network?
no code implementations • 24 Oct 2016 • Mohsen Ghafoorian, Nico Karssemeijer, Tom Heskes, Mayra Bergkamp, Joost Wissink, Jiri Obels, Karlijn Keizer, Frank-Erik de Leeuw, Bram van Ginneken, Elena Marchiori, Bram Platel
In this paper, we propose an automated two-stage method using deep convolutional neural networks (CNN).
no code implementations • 16 Oct 2016 • Mohsen Ghafoorian, Nico Karssemeijer, Tom Heskes, Inge van Uden, Clara Sanchez, Geert Litjens, Frank-Erik de Leeuw, Bram van Ginneken, Elena Marchiori, Bram Platel
The anatomical location of imaging features is of crucial importance for accurate diagnosis in many medical tasks.
We investigate the relation of conductance with weighted kernel k-means for a single community, which leads to the introduction of a new objective function, $\sigma$-conductance.
We argue that this is a desirable property, provide conditions under which NMF quality functions are local, and propose a novel class of local probabilistic NMF quality functions for soft graph clustering.
This motivates the derivation of a new family of quality functions, adaptive scale modularity, which does satisfy the proposed axioms.