Notes

TrueImage: A product Learning Algorithm to further improve the Quality of Telehealth Images.
Finally, in order to realize the task of few-shot HOI, we reorganize 2 HOI benchmark datasets with 2 split strategies, i.e., HICO-NN, TUHOI-NN, HICO-NF, and TUHOI-NF. Extensive experimental results on these datasets have demonstrated the effectiveness of our proposed SAPNet approach.A dynamic model to analyze the thickness-shear vibration of a circular quartz crystal plate with multiple concentric ring electrodes on its upper and bottom surfaces is established with the aid of coordinate transformation. The theoretical solution is obtained, which can be written in a superposition form of Mathieu functions and modified Mathieu functions. The convergence of the solution is demonstrated, and the correctness is numerically validated via results from the finite element method (FEM). Subsequently, a systematic investigation is carried out to quantify the effect of the electrode size on the energy trapping phenomenon, i.e., the resonant frequency and mode shape, which reveals that the ring electrode has a great influence on the work performance of resonators. With the increase of the electrode inertia, i.e., the radius and mass ratio, new trapped modes emergence with the vibration mainly focused on the plate with partial electrodes. Besides, owing to the anisotropy, degenerated trapped modes have different resonant frequencies and the frequency discrepancy between them will become smaller for higher modes. click here Finally, the influence of multiple ring electrodes is investigated, and the qualitative analysis and quantitative results demonstrate that multiple ring electrodes will lead to a more uniform mass sensitivity compared with a single ring electrode. The outcome is widely applicable, which can provide theoretical guidance for the structural design and manufacturing of quartz resonators, as well as a thorough interpretation about the underlying physical mechanism.Transcranial focused ultrasound is a novel noninvasive therapeutic modality for glioblastoma and other disorders of the brain. However, because the phase aberrations caused by the skull need to be corrected with computed tomography (CT) images, the transcranial transducer is tightly fixed on the patient's head to avoid any variation in the relative position, and the focus shifting relies mainly on the capacity for electronic beam steering. Due to the presence of grating lobes and the rapid degradation of the focus quality with increasing focus-shifting distance, transcranial focus-shifting sonication may damage healthy brain tissue unintentionally. To reduce the risks associated with transcranial focused ultrasound therapy, linear frequency-modulated (FM) excitation is proposed. The k-space corrected pseudospectral time domain (PSTD) and acoustic holography approach based on the Rayleigh integral are combined to calculate the distribution of the deposited acoustic power. The corresponding simulation was performed with axial/lateral focus shifting at different distances. The distributions of the deposited acoustic power show that linear FM excitation can effectively suppress undesired prefocal grating lobes without compromising focus quality.Interactive segmentation has recently been explored to effectively and efficiently harvest high-quality segmentation masks by iteratively incorporating user hints. While iterative in nature, most existing interactive segmentation methods tend to ignore the dynamics of successive interactions and take each interaction independently. We here propose to model iterative interactive image segmentation with a Markov decision process (MDP) and solve it with reinforcement learning (RL) where each voxel is treated as an agent. Considering the large exploration space for voxel-wise prediction and the dependence among neighboring voxels for the segmentation tasks, multi-agent reinforcement learning is adopted, where the voxel-level policy is shared among agents. Considering that boundary voxels are more important for segmentation, we further introduce a boundary-aware reward, which consists of a global reward in the form of relative cross-entropy gain, to update the policy in a constrained direction, and a boundary reward in the form of relative weight, to emphasize the correctness of boundary predictions. To combine the advantages of different types of interactions, i.e., simple and efficient for point-clicking, and stable and robust for scribbles, we propose a supervoxel-clicking based interaction design. Experimental results on four benchmark datasets have shown that the proposed method significantly outperforms the state-of-the-arts, with the advantage of fewer interactions, higher accuracy, and enhanced robustness.Capturing the 'mutual gaze' of people is essential for understanding and interpreting the social interactions between them. To this end, this paper addresses the problem of detecting people Looking At Each Other (LAEO) in video sequences. For this purpose, we propose LAEO-Net++, a new deep CNN for determining LAEO in videos. In contrast to previous works, LAEO-Net++ takes spatio-temporal tracks as input and reasons about the whole track. It consists of three branches, one for each character's tracked head and one for their relative position. Moreover, we introduce two new LAEO datasets UCO-LAEO and AVA-LAEO. A thorough experimental evaluation demonstrates the ability of LAEO-Net++ to successfully determine if two people are LAEO and the temporal window where it happens. Our model achieves state-of-the-art results on the existing TVHID-LAEO video dataset, significantly outperforming previous approaches. Finally, we apply LAEO-Net++ to a social network, where we automatically infer the social relationship between pairs of people based on the frequency and duration that they LAEO, and show that LAEO can be a useful tool for guided search of human interactions in videos.We present the lifted proximal operator machine (LPOM) to train fully-connected feed-forward neural networks. LPOM represents the activation function as an equivalent proximal operator and adds the proximal operators to the objective function of a network as penalties. LPOM is block multi-convex in all layer-wise weights and activations. This allows us to develop a new block coordinate descent (BCD) method with convergence guarantee to solve it. Due to the novel formulation and solving method, LPOM only uses the activation function itself and does not require any gradient steps. Thus it avoids the gradient vanishing or exploding issues, which are often blamed in gradient-based methods. Also, it can handle various non-decreasing Lipschitz continuous activation functions. Additionally, LPOM is almost as memory-efficient as stochastic gradient descent and its parameter tuning is relatively easy. We further implement and analyze the parallel solution of LPOM. We first propose a general asynchronous-parallel BCD method with convergence guarantee.
Homepage: https://www.selleckchem.com/products/az-33.html