Kromannpeters5539

Unsupervised domain adaptation (UDA) aims to address the domain-shift problem between a labeled source domain and an unlabeled target domain. Many efforts have been made to eliminate the mismatch between the distributions of training and testing data by learning domain-invariant representations. However, the learned representations are usually not task-oriented, i.e., being class-discriminative and domain-transferable simultaneously. This drawback limits the flexibility of UDA in complicated open-set tasks where no labels are shared between domains. In this paper, we break the concept of task-orientation into task-relevance and task-irrelevance, and propose a dynamic task-oriented disentangling network (DTDN) to learn disentangled representations in an end-to-end fashion for UDA. The dynamic disentangling network effectively disentangles data representations into two components the task-relevant ones embedding critical information associated with the task across domains, and the task-irrelevant ones with the remaining non-transferable or disturbing information. These two components are regularized by a group of task-specific objective functions across domains. KN-93 mw Such regularization explicitly encourages disentangling and avoids the use of generative models or decoders. Experiments in complicated, open-set scenarios (retrieval tasks) and empirical benchmarks (classification tasks) demonstrate that the proposed method captures rich disentangled information and achieves superior performance.In the task of monocular 3D pose estimation, the estimation errors of limb joints (i.e., wrist, ankle, etc) with a higher degree of freedom(DOF) are larger than that of others (i.e., hip, thorax, etc). Specifically, errors may accumulate along the physiological structure of human body parts, and trajectories of joints with higher DOF bring in higher complexity. To address this problem, we propose a limb pose aware framework, involving a kinematic constraint aware network as well as a trajectory aware temporal module, to improve the 3D prediction accuracy of limb joint positions. Two kinematic constraints named relative bone angles and absolute bone angles are introduced in this paper, the former being used for building the angular relation between adjacent bones and the latter for building the angular relation between bones and the camera plane. As a joint result of two constraints, our work suppresses errors accumulated along limbs. Furthermore, we propose a trajectory-aware network, named as Hierarchical Transformer, which takes temporal trajectories of joints as input and generates fused trajectory estimation as a result. The Hierarchical Transformer consists of Transformer Encoder blocks and aims at improving the performance of fusing temporal features. Under the effect of kinematic constraints and trajectory network, we alleviate the problem of errors accumulated along limbs and achieve promising results. Most of the off-the-shelf 2D pose estimators can be easily integrated into our framework. We perform extensive experiments on public datasets and validate the effectiveness of the framework. The ablation studies show the strength of each individual sub-module.In this work, we propose a normalized Tanh activate strategy and a lightweight wide-activate recurrent structure to solve three key challenges of the soft-decoding of near-lossless codes 1. How to add an effective strict constrained peak absolute error (PAE) boundary to the network; 2. An end-to-end solution that is suitable for different quantization steps (compression ratios). 3. Simple structure that favors the GPU and FPGA implementation. To this end, we propose a Wide-activated Recurrent structure with a normalized Tanh activate strategy for Soft-Decoding (WRSD). Experiments demonstrate the effectiveness of the proposed WRSD technique that WRSD outperforms better than the state-of-the-art soft decoders with less than 5% number of parameters, and every computation node of WRSD requires less than 64KB storage for the parameters which can be easily cached by most of the current consumer-level GPUs. Source code is available at https//github.com/dota-109/WRSD.Starting from the seminal work of Fully Convolutional Networks (FCN), there has been significant progress on semantic segmentation. However, deep learning models often require large amounts of pixelwise annotations to train accurate and robust models. Given the prohibitively expensive annotation cost of segmentation masks, we introduce a self-training framework in this paper to leverage pseudo labels generated from unlabeled data. In order to handle the data imbalance problem of semantic segmentation, we propose a centroid sampling strategy to uniformly select training samples from every class within each epoch. We also introduce a fast training schedule to alleviate the computational burden. This enables us to explore the usage of large amounts of pseudo labels. Our Centroid Sampling based Self-Training framework (CSST) achieves state-of-the-art results on Cityscapes and CamVid datasets. On PASCAL VOC 2012 test set, our models trained with the original train set even outperform the same models trained on the much bigger augmented train set. This indicates the effectiveness of CSST when there are fewer annotations. We also demonstrate promising few-shot generalization capability from Cityscapes to BDD100K and from Cityscapes to Mapillary datasets.

To investigate the feasibility of developing an acoustic measurement library for non-invasive trans-rodent skull ultrasonic focusing at high frequency.

A fiber-optic hydrophone (FOH) was positioned at the geometric focus of a spherically-curved phased array (64 elements, 25 mm diameter, 20 mm radius of curvature). Elements were driven sequentially (3.3 MHz driving frequency) and FOH waveforms were recorded with and without intervening ex-vivo rodent skullcaps. Measurements were carried out on 15 skullcaps (Sprague-Dawley rats, 182-209 g) across 3 fixed transmission regions per specimen. An element-wise measurement library of skull-induced phase differences was constructed using mean values across all specimens for each transmission region. Library-based transcranial phase differences were compared with direct FOH-based measurements across 5 additional skullcaps not included in the library.

Library-based phase corrections deviated less from FOH-based trans-skull phase difference values than those calculated for the water-path case, and restored partial transcranial focal quality relative to that recovered using invasive hydrophone-based corrections. Retrospective analysis suggests comparable performance can be obtained using smaller library sizes.

An acoustic measurement library can facilitate non-invasive transcranial aberration correction in rodents at high frequency.

Library-based focusing represents a practical approach for delivering high-frequency ultrasound brain treatments in small animals.

Library-based focusing represents a practical approach for delivering high-frequency ultrasound brain treatments in small animals.Responses of the human brain to different visual stimuli elicit specific patterns in electroencephalography (EEG) signals. It is confirmed that by analyzing these patterns, we can recognize the category of the visited objects. However, high levels of noise and artifacts in EEG signals and the discrepancies between the recorded data from different subjects in visual object recognition task make classification of cognitive states of subjects a serious challenge. In this research, we present a framework for evaluating machine learning and wrapper channel selection algorithms used for classifying single-trial EEG signals recorded in response to photographic stimuli. It is shown that by correctly mapping the entire EEG data space to informative feature spaces (IFS), the performance of the classification methods can improve significantly. Results outperform the state-of-the-art results and confirm efficiency of the proposed feature selection methods in capturing the most informative EEG channels. This can help to achieve high separability of object categories in single-trial visual object recognition task.

Separation and detection of micro-particles or cells from bio-samples by point-of-care (POC) systems are critical for biomedical and healthcare diagnostic applications. Among the microfluidic separation techniques, the acoustophoresis-based microfluidic separation technique has the advantages of label-free, contactless, and good biocompatibility. However, most of the separation techniques are bulky, requiring additional equipment for analysis, not suitable for POC-based in-field real-time applications. Therefore, we proposed a platform, which integrates an acoustophoresis-based separation device and a lensless imaging sensor into a compact standalone system to solve the problem.

In this system, Standing Surface Acoustic Wave (SSAW) is utilized for label-free particle separation, while lensless imaging is employed for seamless particle detection and counting using self-developed dual-threshold motion detection algorithms. In particular, the microfluidic channel and interdigital transducers (IDTs) were specnd the design methodologies offer a promising POC solution for label-free cell separation and detection in biomedical diagnostics.Acute aortic syndrome (AAS) is classically attributed to three underlying pathologic conditions-aortic dissection (AD), intramural hematoma (IMH), and penetrating atherosclerotic ulcer (PAU). In the majority of cases, the basics of image interpretation are not difficult and have been extensively reviewed in the literature. In this article, the authors extend existing imaging overviews of AAS by highlighting additional factors related to the diagnosis, classification, and characterization of difficult AAS cases. It has been well documented that AAS is caused not only by an AD but by a spectrum of lesions that often have overlap in imaging features and are not clearly distinguishable. Specifically, phase of contrast enhancement, flow artifacts, and flapless AD equivalents can complicate diagnosis and are discussed. While the A/B dichotomy of the Stanford system is still used, the authors subsequently emphasize the Society for Vascular Surgery's new guidelines for the description of acute aortic pathologic conditions given the expanded use of endovascular techniques used in aortic repair. In the final section, atypical aortic rupture and pitfalls are described. As examples of pericardial and shared sheath rupture become more prevalent in the literature, it is important to recognize contrast material third-spacing and mediastinal blood as potential mimics. By understanding these factors related to difficult cases of AAS, the diagnostic radiologist will be able to accurately refine CT interpretation and thus provide information that is best suited to directing management. Online supplemental material is available for this article. ©RSNA, 2021.Injury of the scapholunate ligament (SLL) complex can lead to scapholunate dissociation, characterized by scapholunate interval widening and volar rotary subluxation of the scaphoid. Loss of the mechanical linkage between the scaphoid and lunate results in carpal instability and eventual scapholunate advanced collapse (SLAC) arthropathy. SLL complex injuries vary from acute and traumatic to chronic and degenerative. A staging system can be used to guide treatment options for these injuries on the basis of the reparability of the SLL dorsal band, carpal alignment and malalignment reducibility, and cartilage damage. Preoperative imaging with radiography and MRI is a component of injury staging and aids in planning surgical procedures. If the SLL dorsal band is reparable, then direct primary ligament repair with dorsal capsulodesis or dorsal intercarpal (DIC) ligament transfer can be performed. If the SLL dorsal band is irreparable with normal alignment or reducible malalignment, then reconstruction can be performed.