Explore the vast range of titles available.

Aliasing refers to the phenomenon that high frequency signals degenerate into completely different ones after sampling. It arises as a problem in the context of deep learning as downsampling layers are widely adopted in deep architectures to reduce parameters and computation. The standard solution is to apply a low-pass filter (e.g., Gaussian blur) before downsampling (Zhang in: ICML, 2020). However, it can be suboptimal to apply the same filter across the entire content, as the frequency of feature maps can vary across both spatial locations and feature channels. To tackle this, we propose an adaptive content-aware low-pass filtering layer, which predicts separate filter weights for each spatial location and channel group of the input feature maps. We investigate the effectiveness and generalization of the proposed method across multiple tasks, including image classification, semantic segmentation, instance segmentation, video instance segmentation, and image-to-image translation. Both qualitative and quantitative results demonstrate that our approach effectively adapts to the different feature frequencies to avoid aliasing while preserving useful information for recognition. Code is available at https://maureenzou.github.io/ddac/.

We evaluate residual aliasing among simultaneously excited and acquired slices in slice accelerated multiband (MB) echo planar imaging (EPI). No in-plane accelerations were used in order to maximize and evaluate achievable slice acceleration factors at 3T. We propose a novel leakage (L-) factor to quantify the effects of signal leakage between simultaneously acquired slices. With a standard 32-channel receiver coil at 3T, we demonstrate that slice acceleration factors of up to eight (MB=8) with blipped controlled aliasing in parallel imaging (CAIPI), in the absence of in-plane accelerations, can be used routinely with acceptable image quality and integrity for whole brain imaging. Spectral analyses of single-shot fMRI time series demonstrate that temporal fluctuations due to both neuronal and physiological sources were distinguishable and comparable up to slice-acceleration factors of nine (MB=9). The increased temporal efficiency could be employed to achieve, within a given acquisition period, higher spatial resolution, increased fMRI statistical power, multiple TEs, faster sampling of temporal events in a resting state fMRI time series, increased sampling of q-space in diffusion imaging, or more quiet time during a scan. [Display omitted] •High slice accelerations using multiband (MB) GRE-EPI with blipped CAIPI.•Acceptable MB factors up to 8 with a 32-channel receiver coil at 3T.•Neuronal and physiological sources are distinguishable at high MB factors.•Leakage (L-) factor evaluates residual aliasing among simultaneously acquired slices.•High temporal efficiency with MB-EPI benefits various applications.

High-performance electric field sensors (EFS) provide support for detecting electric field (E-field) in atmospheric and power grid applications. This paper proposed a method using weakly coupled fork microstructure to enhance the quality factor (Q-factor) and mitigate mode aliasing. Compared to traditional micro electric field mill sensors (MEFM), this approach shows improvements in these performance metrics. The performance of this structure was analysed theoretically and simulated numerically, and a MEMS chip prototype was fabricated. At an air pressure of 0.001 Pa, the sensor achieved a Q factor of 42,423, with a measurement range of 90 kV/m and a resolution of 32 V/m under these conditions. The sensor demonstrates better potential to meet the requirements in atmospheric and power grid applications.

For the majority of run sizes N where N <= 20, the literature reports the best D- and A-optimal designs for the main-effects model which sequentially minimizes the aliasing between main effects and interaction effects and among interaction effects. The only series of run sizes for which all the minimally aliased D- and A-optimal main-effects designs remain unknown are those with run sizes three more than a multiple of four. To address this, in our paper, we propose an algorithm to generate all non-isomorphic D- and A-optimal main-effects designs for run sizes three more than a multiple of four. We enumerate all such designs for run sizes up to 19, report the numbers of designs we obtained, and identify those that minimize the aliasing between main effects and interaction effects and among interaction effects.

Convolutional neural networks leverage their efficient ability to extract common features of images, playing a crucial role in numerous computer vision tasks. Key details such as edges and textures in images often present themselves in the form of high-frequency components, which contain rich semantic information and are essential for accurate image recognition and understanding. However, during the downsampling process, these high-frequency components are improperly mapped to low-frequency components, leading to signal aliasing. This aliasing results in the loss of image detail information and blurred features, significantly affecting the precise extraction of image features by convolutional neural networks and ultimately reducing the performance of the model in various tasks. To effectively address this challenge, this study innovatively proposes the Gaussian Filter Convolution (GFC) module. This module ingeniously utilizes convolution kernels with filtering functions, which can specifically suppress the high-frequency components in the image, reducing the occurrence of signal aliasing at its source, thereby significantly alleviating the aliasing artifacts generated during downsampling. Experimental data show that the model integrated with GFC has significant improvements in key indicators such as model accuracy.

Since the mid-1990s, a series of FES (finite element solution) global ocean tidal atlases has been produced and released with the primary objective to provide altimetry missions with tidal de-aliasing correction at the best possible accuracy. We describe the underlying hydrodynamic and data assimilation design and accuracy assessments for the latest FES2014 release (finalized in early 2016), especially for the altimetry de-aliasing purposes. The FES2014 atlas shows extremely significant improvements compared to the standard FES2004 and (intermediary) FES2012 atlases, in all ocean compartments, especially in shelf and coastal seas, thanks to the unstructured grid flexible resolution, recent progress in the (prior to assimilation) hydrodynamic tidal solutions, and use of ensemble data assimilation technique. Compared to earlier releases, the available tidal constituent's spectrum has been significantly extended, the overall resolution has been augmented, and additional scientific byproducts such as loading and self-attraction, energy diagnostics, or lowest astronomical tides have been derived from the atlas and are available. Compared to the other available global ocean tidal atlases, FES2014 clearly shows improved de-aliasing performance in most of the global ocean areas and has consequently been integrated in satellite altimetry geophysical data records (GDRs) and gravimetric data processing and adopted in recently renewed ITRF standards (International Terrestrial Reference System, 2020). It also provides very accurate open-boundary tidal conditions for regional and coastal modelling.

Addressing the limitations of traditional underdetermined blind separation methods for time-frequency aliasing signal reconstruction, which often require high signal sparsity and yield low reconstruction accuracy, this paper introduces a signal reconstruction based on minimizing non-principal diagonal proportion approach. Based on the matrix diagonalization method, this method makes full use of the independence between source signals, evaluates the diagonalization degree of the covariance matrix by minimizing the proportion of non-principal diagonal elements, and realizes the accurate identification of the mixed vector matrix of each time-frequency point. The results of the simulation indicate that the method put forward attains greater accuracy in reconstruction when contrasted with various traditional techniques in blind separation situations marked by significant underdetermined time-frequency aliasing.

For three-level regular designs, the confounding from the perspectives of both factor and component effects leads to different results. The aliasing properties of factor effects are more significant than the latter in the experimental model. In this paper, a new three-level aliasing pattern is proposed to evaluate the degree of aliasing among different factors. Based on the classification pattern, a new criterion is introduced for choosing optimal three-level regular designs. Then, we analyze the relationship between the criterion and the existing criteria, including general minimum lower-order confounding, entropy, minimum aberration, and clear effects. The results show that the classification patterns of other criteria can be expressed as functions of our proposed pattern. Further, an aliasing algorithm is provided, and all 27-run, some of the 81-run, and 243-run three-level designs are listed in tables and compared with the rankings under other criteria. A real example is provided to illustrate the proposed methods.

The ability to present three-dimensional (3D) scenes with continuous depth sensation has a profound impact on virtual and augmented reality, human–computer interaction, education and training. Computer-generated holography (CGH) enables high-spatio-angular-resolution 3D projection via numerical simulation of diffraction and interference 1 . Yet, existing physically based methods fail to produce holograms with both per-pixel focal control and accurate occlusion 2 , 3 . The computationally taxing Fresnel diffraction simulation further places an explicit trade-off between image quality and runtime, making dynamic holography impractical 4 . Here we demonstrate a deep-learning-based CGH pipeline capable of synthesizing a photorealistic colour 3D hologram from a single RGB-depth image in real time. Our convolutional neural network (CNN) is extremely memory efficient (below 620 kilobytes) and runs at 60 hertz for a resolution of 1,920 × 1,080 pixels on a single consumer-grade graphics processing unit. Leveraging low-power on-device artificial intelligence acceleration chips, our CNN also runs interactively on mobile (iPhone 11 Pro at 1.1 hertz) and edge (Google Edge TPU at 2.0 hertz) devices, promising real-time performance in future-generation virtual and augmented-reality mobile headsets. We enable this pipeline by introducing a large-scale CGH dataset (MIT-CGH-4K) with 4,000 pairs of RGB-depth images and corresponding 3D holograms. Our CNN is trained with differentiable wave-based loss functions 5 and physically approximates Fresnel diffraction. With an anti-aliasing phase-only encoding method, we experimentally demonstrate speckle-free, natural-looking, high-resolution 3D holograms. Our learning-based approach and the Fresnel hologram dataset will help to unlock the full potential of holography and enable applications in metasurface design 6 , 7 , optical and acoustic tweezer-based microscopic manipulation 8 – 10 , holographic microscopy 11 and single-exposure volumetric 3D printing 12 , 13 . A deep-learning-based approach using a convolutional neural network is used to synthesize photorealistic colour three-dimensional holograms from a single RGB-depth image in real time, and termed tensor holography.

Cache attacks exploit memory access patterns of cryptographic implementations. Constant-time implementation techniques have become an indispensable tool in fighting cache timing attacks. These techniques engineer the memory accesses of cryptographic operations to follow a uniform key independent pattern. However, the constant-time behavior is dependent on the underlying architecture, which can be highly complex and often incorporates unpublished features. The CacheBleed attack targets cache bank conflicts and thereby invalidates the assumption that microarchitectural side-channel adversaries can only observe memory with cache line granularity. In this work, we propose MemJam, which utilizes 4K Aliasing to establish a side-channel attack that exploits false dependency of memory read-after-write events and provides a high quality intra cache line timing channel. As a proof of concept, we demonstrate the first key recovery attacks on constant-time implementations of all symmetric block ciphers supported in the current intel integrated performance primitives (Intel IPP) cryptographic library: triple DES, AES and SM4. Further, we demonstrate the first intra cache level timing attack on SGX by reproducing the AES key recovery results on an enclave that performs encryption using the aforementioned constant-time implementation of AES. Our results show that we can not only use this side channel to efficiently attack memory dependent cryptographic operations but also to bypass proposed protections. Compared to CacheBleed, which is limited to older processor generations, MemJam is the first intra cache level attack applicable to all major Intel processors including the latest generations and also applies to the SGX extension.