1. |
- Georgescu, Mariana-Iuliana, et al.
(författare)
-
Multimodal Multi-Head Convolutional Attention with Various Kernel Sizes for Medical Image Super-Resolution
- 2023
-
Ingår i: 2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV). - : IEEE COMPUTER SOC. - 9781665493468 - 9781665493475 ; , s. 2194-2204
-
Konferensbidrag (refereegranskat)abstract
- Super-resolving medical images can help physicians in providing more accurate diagnostics. In many situations, computed tomography (CT) or magnetic resonance imaging (MRI) techniques capture several scans (modes) during a single investigation, which can jointly be used (in a multimodal fashion) to further boost the quality of super-resolution results. To this end, we propose a novel multimodal multi-head convolutional attention module to super-resolve CT and MRI scans. Our attention module uses the convolution operation to perform joint spatial-channel attention on multiple concatenated input tensors, where the kernel (receptive field) size controls the reduction rate of the spatial attention, and the number of convolutional filters controls the reduction rate of the channel attention, respectively. We introduce multiple attention heads, each head having a distinct receptive field size corresponding to a particular reduction rate for the spatial attention. We integrate our multimodal multi-head convolutional attention (MMHCA) into two deep neural architectures for super-resolution and conduct experiments on three data sets. Our empirical results show the superiority of our attention module over the state-of-the-art attention mechanisms used in super-resolution. Moreover, we conduct an ablation study to assess the impact of the components involved in our attention module, e.g. the number of inputs or the number of heads. Our code is freely available at https://github.com/lilygeorgescu/MHCA.
|
|
2. |
- Kumagai, Yoshiaki, et al.
(författare)
-
Suppression of thermal nanoplasma emission in clusters strongly ionized by hard x-rays
- 2021
-
Ingår i: Journal of Physics B. - : Institute of Physics Publishing (IOPP). - 0953-4075 .- 1361-6455. ; 54:4
-
Tidskriftsartikel (refereegranskat)abstract
- Using electron and ion spectroscopy, we studied the electron and nuclear dynamics in similar to 50 000-atom large krypton clusters, following excitation with an intense hard x-ray pulse. Beyond the single pulse experiment, we also present the results of a time-resolved, x-ray pump-near-infrared probe measurement that allows one to learn about the time evolution of the system. After core ionization of the atoms by x-ray photons, trapped Auger and secondary electrons form a nanoplasma in which the krypton ions are embedded, according to the already published scenario. While the ion data show expected features, the electron emission spectra miss the expected pump-probe delay-dependent enhancement except for a slight enhancement in the energy range below 2 eV. Theoretical simulations help to reveal that, due to the deep trapping potential of the ions during the long time expansion accompanied by electron-ion recombination, thermal emission from the transient nanoplasma becomes quenched.
|
|
3. |
- Nandi, Saikat, et al.
(författare)
-
Generation of entanglement using a short-wavelength seeded free-electron laser
- 2024
-
Ingår i: SCIENCE ADVANCES. - 2375-2548. ; 10:16
-
Tidskriftsartikel (refereegranskat)abstract
- Quantum entanglement between the degrees of freedom encountered in the classical world is challenging to observe due to the surrounding environment. To elucidate this issue, we investigate the entanglement generated over ultrafast timescales in a bipartite quantum system comprising two massive particles: a free-moving photoelectron, which expands to a mesoscopic length scale, and a light-dressed atomic ion, which represents a hybrid state of light and matter. Although the photoelectron spectra are measured classically, the entanglement allows us to reveal information about the dressed-state dynamics of the ion and the femtosecond extreme ultraviolet pulses delivered by a seeded free-electron laser. The observed generation of entanglement is interpreted using the time-dependent von Neumann entropy. Our results unveil the potential for using short-wavelength coherent light pulses from free-electron lasers to generate entangled photoelectron and ion systems for studying spooky action at a distance.
|
|
4. |
- Nandi, Saikat, et al.
(författare)
-
Unraveling Rabi dynamics with a seeded FEL at XUV wavelength
- 2023
-
Ingår i: 2023 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2023. - 9798350345995
-
Konferensbidrag (refereegranskat)abstract
- Rabi oscillations, a prominent feature of coherent light-matter interaction arise when a two-level system interacts periodically with an external electromagnetic field [1]. Despite being a cornerstone in quantum physics, they are usually studied in the long-wavelength region, ranging from mid-infrared to visible. Here, we demonstrate that intense femtosecond extreme-ultraviolet (XUV) pulses from FERMI seeded free-electron laser [2] can drive Rabi oscillations between the two levels: 1s2 and 1s4p in helium.
|
|
5. |
- Ristea, Nicolae-Catalin, et al.
(författare)
-
CyTran: A cycle-consistent transformer with multi-level consistency for non-contrast to contrast CT translation
- 2023
-
Ingår i: Neurocomputing. - : ELSEVIER. - 0925-2312 .- 1872-8286. ; 538
-
Tidskriftsartikel (refereegranskat)abstract
- We propose a novel approach to translate unpaired contrast computed tomography (CT) scans to noncontrast CT scans and the other way around. Solving this task has two important applications: (i) to automatically generate contrast CT scans for patients for whom injecting contrast substance is not an option, and (ii) to enhance the alignment between contrast and non-contrast CT by reducing the differences induced by the contrast substance before registration.Our approach is based on cycle-consistent generative adversarial convolutional transformers, for short, CyTran. Our neural model can be trained on unpaired images, due to the integration of a multi-level cycleconsistency loss. Aside from the standard cycle-consistency loss applied at the image level, we propose to apply additional cycle-consistency losses between intermediate feature representations, which enforces the model to be cycle-consistent at multiple representations levels, leading to superior results. To deal with high-resolution images, we design a hybrid architecture based on convolutional and multi-head attention layers. In addition, we introduce a novel data set, Coltea-Lung-CT-100W, containing 100 3D triphasic lung CT scans (with a total of 37,290 images) collected from 100 female patients (there is one examination per patient). Each scan contains three phases (non-contrast, early portal venous, and late arterial), allowing us to perform experiments to compare our novel approach with state-of-the-art methods for image style transfer.Our empirical results show that CyTran outperforms all competing methods. Moreover, we show that CyTran can be employed as a preliminary step to improve a state-of-the-art medical image alignment method. We release our novel model and data set as open source at: https://github.com/ristea/cycletransformer.Our qualitative and subjective human evaluations reveal that CyTran is the only approach that does not introduce visual artifacts during the translation process. We believe this is a key advantage in our application domain, where medical images need to precisely represent the scanned body parts. (c) 2023 Elsevier B.V. All rights reserved.
|
|