The Experts below are selected from a list of 9414 Experts worldwide ranked by ideXlab platform
Bernhard Kainz - One of the best experts on this subject based on the ideXlab platform.
-
deepcut object segmentation from bounding box annotations using convolutional neural networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passeratpalmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled weak annotations, in our case bounding boxes. It extends the approach of the well-known GrabCut[1] method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.
-
DeepCut: Object Segmentation from Bounding Box Annotations Using Convolutional Neural Networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passerat-palmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled with bounding box annotations. It extends the approach of the well-known GrabCut method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.
Martin Rajchl - One of the best experts on this subject based on the ideXlab platform.
-
deepcut object segmentation from bounding box annotations using convolutional neural networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passeratpalmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled weak annotations, in our case bounding boxes. It extends the approach of the well-known GrabCut[1] method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.
-
DeepCut: Object Segmentation from Bounding Box Annotations Using Convolutional Neural Networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passerat-palmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled with bounding box annotations. It extends the approach of the well-known GrabCut method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.
Trixie Wagner - One of the best experts on this subject based on the ideXlab platform.
-
structural analysis of metastable pharmaceutical loratadine form ii by 3d electron diffraction and dft d Energy Minimisation
CrystEngComm, 2020Co-Authors: Grahame Woollam, Jacco Van De Streek, Partha P Das, Enrico Mugnaioli, Iryna Andrusenko, Athanassios S Galanis, S Nicolopoulos, Mauro Gemmi, Trixie WagnerAbstract:Metastable polymorphs typically display higher solubility than their thermodynamically stable counterparts, whilst having dissimilar mechanical and biopharmaceutical properties. It is unsurprising then, that generic and innovator companies alike pursue isolation and characterisation of these materials. Here we report the determination of the crystal structure of the metastable form II of loratadine using a combination of low-resolution 3D electron diffraction data and density functional theory. Importantly, electron diffraction was able to establish that the crystallites were phase pure i.e. no other polymorphic forms were identified throughout the sample. 3D data collected at room temperature circumvented potential phase changes, conveniently preserving the metastable polymorph during structural elucidation. The limited resolution of the electron diffraction data (>1 A), combined with the complexity of configurational disorder and possible beam-induced amorphization, meant that the structure could not be obtained by ab initio direct methods. This is a recurrent situation for nanocrystalline pharmaceutical crystals. Instead, two possible starting models arose from simulated annealing based on diffraction data alone. Density functional theory Energy Minimisation followed, determining the correct model, with an independent validation of the experimental structural solution, comparing favourably to single-crystal X-ray diffraction studies. Our results reveal a promising protocol enabling the exploitation of electron diffraction data with limited resolution obtained from beam sensitive organic materials. The method is widely extendable to a number of pharmaceutical compounds that are not amenable for the growth of large single crystals required by X-ray diffraction, and for which efficient structure determination is required.
Jacco Van De Streek - One of the best experts on this subject based on the ideXlab platform.
-
structural analysis of metastable pharmaceutical loratadine form ii by 3d electron diffraction and dft d Energy Minimisation
CrystEngComm, 2020Co-Authors: Grahame Woollam, Jacco Van De Streek, Partha P Das, Enrico Mugnaioli, Iryna Andrusenko, Athanassios S Galanis, S Nicolopoulos, Mauro Gemmi, Trixie WagnerAbstract:Metastable polymorphs typically display higher solubility than their thermodynamically stable counterparts, whilst having dissimilar mechanical and biopharmaceutical properties. It is unsurprising then, that generic and innovator companies alike pursue isolation and characterisation of these materials. Here we report the determination of the crystal structure of the metastable form II of loratadine using a combination of low-resolution 3D electron diffraction data and density functional theory. Importantly, electron diffraction was able to establish that the crystallites were phase pure i.e. no other polymorphic forms were identified throughout the sample. 3D data collected at room temperature circumvented potential phase changes, conveniently preserving the metastable polymorph during structural elucidation. The limited resolution of the electron diffraction data (>1 A), combined with the complexity of configurational disorder and possible beam-induced amorphization, meant that the structure could not be obtained by ab initio direct methods. This is a recurrent situation for nanocrystalline pharmaceutical crystals. Instead, two possible starting models arose from simulated annealing based on diffraction data alone. Density functional theory Energy Minimisation followed, determining the correct model, with an independent validation of the experimental structural solution, comparing favourably to single-crystal X-ray diffraction studies. Our results reveal a promising protocol enabling the exploitation of electron diffraction data with limited resolution obtained from beam sensitive organic materials. The method is widely extendable to a number of pharmaceutical compounds that are not amenable for the growth of large single crystals required by X-ray diffraction, and for which efficient structure determination is required.
-
Structural Analysis of Metastable Pharmaceutical Loratadine Form II, by 3D Electron Diffraction and DFT+D Energy Minimisation
'Royal Society of Chemistry (RSC)', 2020Co-Authors: Woollam Grahame, Jacco Van De Streek, Das, Partha P., Mugnaioli Enrico, Andrusenko Iryna, Galanis, Athanassios S., Nicolopoulos Stavros, Gemmi Mauro, Wagner BeatrixAbstract:Metastable polymorphs typically display higher solubility than their thermodynamically stable counterparts, whilst having dissimilar mechanical and biopharmaceutical properties. It is unsurprising then, that generic and innovator companies alike pursue isolation and characterisation of these materials. Here we report the determination of the crystal structure of the metastable Form II of loratadine using a combination of low-resolution 3D electron diffraction data and density functional theory. Importantly, electron diffraction was able to establish that the crystallites were phase pure i.e. no other polymorphic forms were identified throughout the sample. 3D data collected at room temperature circumvented potential phase changes, conveniently preserving the metastable polymorph during structural elucidation. The limited resolution of the electron diffraction data (> 1 Å), combined with the complexity of configurational disorder and possible beam-induced amorphization, meant that the structure could not be obtained by ab-initio direct methods. This is a recurrent situation for nanocrystalline pharmaceutical crystals. Instead, two possible starting models arose from simulated annealing based on diffraction data alone. Density Functional Theory Energy Minimisation followed, determining the correct model, with an independent validation of the experimental structural solution, comparing favourably to single-crystal X-ray diffraction studies. Our results reveal a promising protocol enabling the exploitation of electron diffraction data with limited resolution obtained from beam sensitive organic materials. The method is widely extendable to a number of pharmaceutical compounds that are not amenable for the growth of large single crystals required by X-ray diffraction, and for which an efficient structure determination is required
-
experimental verification of a subtle low temperature phase transition suggested by dft d Energy Minimisation
CrystEngComm, 2011Co-Authors: Andrew D Bond, Katarzyna A Solanko, Jacco Van De Streek, Marcus A NeumannAbstract:Energy Minimisation of 2,6-bis(2,4-dichlorobenzylidene)cyclohexanone using dispersion-corrected Density Functional Theory (DFT-D) suggested a subtle low-temperature phase transition that has been verified by experiment. The results demonstrate that DFT-D calculations are sufficiently reliable to guide experimental studies towards targets most likely to exhibit interesting temperature-dependent structural change.
Joseph V Hajnal - One of the best experts on this subject based on the ideXlab platform.
-
deepcut object segmentation from bounding box annotations using convolutional neural networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passeratpalmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled weak annotations, in our case bounding boxes. It extends the approach of the well-known GrabCut[1] method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.
-
DeepCut: Object Segmentation from Bounding Box Annotations Using Convolutional Neural Networks
IEEE Transactions on Medical Imaging, 2017Co-Authors: Martin Rajchl, Matthew C.h. Lee, Ozan Oktay, Wenjia Bai, Konstantinos Kamnitsas, Mary A. Rutherford, Mellisa Damodaram, Joseph V Hajnal, Jonathan Passerat-palmbach, Bernhard KainzAbstract:In this paper, we propose DeepCut, a method to obtain pixelwise object segmentations given an image dataset labelled with bounding box annotations. It extends the approach of the well-known GrabCut method to include machine learning by training a neural network classifier from bounding box annotations. We formulate the problem as an Energy Minimisation problem over a densely-connected conditional random field and iteratively update the training targets to obtain pixelwise object segmentations. Additionally, we propose variants of the DeepCut method and compare those to a naive approach to CNN training under weak supervision. We test its applicability to solve brain and lung segmentation problems on a challenging fetal magnetic resonance dataset and obtain encouraging results in terms of accuracy.