Accuracy and consistency are two key factors in computer-assisted magnetic resonance (MR) image analysis. However, contrast variation from site to site caused by lack of standardization in MR acquisition impedes consistent measurements. In recent yea
rs, image harmonization approaches have been proposed to compensate for contrast variation in MR images. Current harmonization approaches either require cross-site traveling subjects for supervised training or heavily rely on site-specific harmonization models to encourage harmonization accuracy. These requirements potentially limit the application of current harmonization methods in large-scale multi-site studies. In this work, we propose an unsupervised MR harmonization framework, CALAMITI (Contrast Anatomy Learning and Analysis for MR Intensity Translation and Integration), based on information bottleneck theory. CALAMITI learns a disentangled latent space using a unified structure for multi-site harmonization without the need for traveling subjects. Our model is also able to adapt itself to harmonize MR images from a new site with fine tuning solely on images from the new site. Both qualitative and quantitative results show that the proposed method achieves superior performance compared with other unsupervised harmonization approaches.
TaAs and NbAs are two of the earliest identified Weyl semimetals that possess many intriguing optical properties, such as chirality-dependent optical excitations and giant second harmonic generation (SHG). Linear and nonlinear optics have been employ
ed as tools to probe the Weyl physics in these crystals. Here we extend these studies to address two important points: determining the complete anisotropic dielectric response, and to explore if and how they can reveal essential Weyl physics. For the first time, we determine the complete anisotropic dielectric functions of TaAs and NbAs by combining spectroscopic ellipsometry and density functional theory (DFT). Parameterized Lorentz oscillators are reported from 1.2-6 eV (experiment) and 0-6 eV (DFT), and good agreement is shown between them. Both linear and nonlinear optical properties have been reported to reveal Weyl physics. We suggest that strong optical resonances from trivial bands are the likely origin of the large optical second harmonic generation previously reported at these energies. Furthermore, by comparing the contribution of a small k-space centered around the Weyl cones to the total linear dielectric function, we find that these contributions are highly anisotropic and are <25% of the total dielectric function below 0.5 eV; above 1eV, these contributions are negligible. Thus, the study of Weyl physics using optical techniques requires very low energies and even there, a careful assessment is required in distinguishing the much smaller contributions of the Weyl bands from the dominant contributions of the trivial bands and Drude response to the total dielectric function.
Domain shift is a major problem for deploying deep networks in clinical practice. Network performance drops significantly with (target) images obtained differently than its (source) training data. Due to a lack of target label data, most work has foc
used on unsupervised domain adaptation (UDA). Current UDA methods need both source and target data to train models which perform image translation (harmonization) or learn domain-invariant features. However, training a model for each target domain is time consuming and computationally expensive, even infeasible when target domain data are scarce or source data are unavailable due to data privacy. In this paper, we propose a novel self domain adapted network (SDA-Net) that can rapidly adapt itself to a single test subject at the testing stage, without using extra data or training a UDA model. The SDA-Net consists of three parts: adaptors, task model, and auto-encoders. The latter two are pre-trained offline on labeled source images. The task model performs tasks like synthesis, segmentation, or classification, which may suffer from the domain shift problem. At the testing stage, the adaptors are trained to transform the input test image and features to reduce the domain shift as measured by the auto-encoders, and thus perform domain adaptation. We validated our method on retinal layer segmentation from different OCT scanners and T1 to T2 synthesis with T1 from different MRI scanners and with different imaging parameters. Results show that our SDA-Net, with a single test subject and a short amount of time for self adaptation at the testing stage, can achieve significant improvements.
