Structure-from-Motion (SfM) is a fundamental 3D vision task for recovering camera parameters and scene geometry from multi-view images. While recent deep learning advances enable accurate Monocular Depth Estimation (MDE) from single images without depending on camera motion, integrating MDE into SfM remains a challenge. Unlike conventional triangulated sparse point clouds, MDE produces dense depth maps with significantly higher error variance. Inspired by modern RANSAC estimators, we propose Marginalized Bundle Adjustment (MBA) to mitigate MDE error variance leveraging its density. With MBA, we show that MDE depth maps are sufficiently accurate to yield SoTA or competitive results in SfM and camera relocalization tasks. Through extensive evaluations, we demonstrate consistently robust performance across varying scales, ranging from few-frame setups to large multiview systems with thousands of images. Our method highlights the significant potential of MDE in multi-view 3D vision. Code is available at https://marginalizedba.github.io/.

Hybrid beamforming architectures provide a trade-off between beamforming performance, power consumption, and hardware cost, rendering them suitable for 5G millimeter-wave (mmWave) communications. However, the limited number of radio frequency (RF) chains compared to the large number of antenna elements presents challenges for high-resolution direction of arrival (DoA) and direction of departure (DoD) estimation. While methods like compressive sensing (CS) and beamspace processing have been explored, they suffer from random phase errors and a high computational complexity. In this paper, we propose an ESPRIT-based 2-dimensional (2-D) phase noise parameter compensation (PNPC) algorithm for joint DoD and DoA estimation of the dominant multipath components. The algorithm extends the 2-D beamspace shift invariance equations and introduces a new structure, where the phase error vector resides in the nullspace of a combined matrix. This enables the joint estimation of DoDs, DoAs, and phase errors through an iterative procedure. Our simulation results confirm that the proposed 2-D PNPC-DFT-ESPRIT algorithm achieves a performance close to the ideal 2-D DFT-ESPRIT algorithm without phase errors, even in the challenging case of single RF chains on the transmit and the receive sides.

The widespread use of gradient-based optimization has motivated the adaptation of various classical algorithms into differentiable solvers compatible with learning pipelines. In this paper, we investigate the enhancement of traditional geometric query problems such that the result consists of both the geometric function as well as its gradient. Specifically, we study the fundamental problem of distance queries against a set of points P in ℝd, which also underlies various similarity measures for learning algorithms.

Accurate estimation and compensation of the carrier frequency offset (CFO) are crucial for ensuring reliable OFDM communications in high mobility scenarios. Paired tracking reference symbols are defined by 5G cellular systems to allow low-complexity CFO estimation for a terminal device. However, estimation ambiguity can occur when the actual CFO exceeds the maximal capture range, which is determined by the time interval between the paired tracking reference symbols. This paper proposes a low-complexity method to resolve the CFO ambiguity issue using a leakage ratio (LR) function. The technique can significantly increase the effective CFO capture range without changing the time pattern of the tracking symbols. Our simulations also show that the proposed method can reliably classify the correct CFO candidate even in low SNR conditions.

This paper presents a self-supervised learning approach for online beam synthesis using the reference signal received power (RSRP) measurements in mmWave band communications. We propose a sparse structure which can effectively approximate the true antenna space covariance matrix (ASCM) of the mmWave channel with only a few effective paths. The sparse structure is then forced within the latent space of a model-aided autoencoder (MAE), whose encoding part is a neural network (NN) which estimates the parameters of the effective ASCM by the measured RSRPs, while the decoding part is a closed-form model which reconstructs the RSRPs from the estimated effective ASCM. The MAE is trained based on the similarity loss between the measured RSRPs and the reconstructed RSRPs, such that the ground truth channel state information (CSI) is not needed during the training. The predicted effective spatial covariance matrix is then used to compute the optimal communication beam. Simulations show that the proposed scheme can significantly improve the beamforming performance especially in non-line-of-slight (NLOS) channel conditions.

BACKGROUND: Detecting biomarkers is a key objective in microbiome research, often done through 16S rRNA amplicon sequencing or shotgun metagenomic analysis. A critical step in this process is differential abundance (DA) analysis, which aims to pinpoint taxa whose abundance significantly differs between groups. However, DA analysis remains challenging due to high dimensionality, compositionality, sparsity, inter-taxa correlations, uneven abundance distributions, and missing values-all which hinder our ability to model the data accurately. Despite the availability of many DA tools, balancing high statistical power with effective false discovery rate (FDR) control remains a major limitation. RESULTS: Here, we introduce a novel approach for DA analysis that integrates counts adjusted with Trimmed Mean of M-values (CTF) normalization and Centered Log Ratio (CLR) transformation with Generalized Estimating Equation (GEE) model. We benchmarked our approach against eight widely used tools employing both simulated and real datasets in cross-sectional and longitudinal settings. While several tools (e.g. MetagenomeSeq, edgeR, DESeq2 and Lefse) achieved high sensitivity, they often failed to adequately control the FDR. In contrast, our method demonstrated high sensitivity and specificity when compared to other approaches that successfully controlled the FDR, including ALDEx2, limma-voom, ANCOM, and ANCOM-BC2. CONCLUSIONS: Our approach effectively addresses key challenges in microbiome data analysis across both cross-sectional and longitudinal designs. Integrated into the R package metaGEENOME (https://github.com/M-Mysara/metaGEENOME), our framework provides a flexible, scalable and statistically robust solution for DA analysis, offering improved FDR control and enhanced performance for biomarker discovery in microbiome studies.

Estimation of signal parameters via rotational invariant techniques (ESPRIT) based high-resolution parameter estimation algorithms in discrete Fourier transform (DFT) beamspace are efficient gridless schemes for hybrid beamforming architectures. Since the number of coherent DFT measurements is equal to the number of radio frequency (RF) chains that is much smaller than the number of antennas, only a small spatial region can be scanned coherently. The number of coherent measurements can be increased by generating virtual RF chains through time-domain-multiplexed (TDM) measurements. However, due to hardware imperfections, random phase jump errors between the TDM measurements degrade the direction of arrival (DoA) estimation accuracy. This paper proposes a new 2-dimensional (2-D) beamspace phase noise parameter compensation (PNPC) algorithm, called 2-D PNPC-DFT-ESPRIT, which jointly estimates the random phase jump errors and the DoAs for uniform rectangular arrays (URAs). The 2-D PNPC-DFT-ESPRIT algorithm modifies the beamspace shift invariance equations for URAs, which consider the contributions of the random phase jump errors and the DoAs. An iterative procedure is developed to estimate both jointly. The simulations show that, even in the presence of random phase jump errors, the 2-D PNPC-DFT-ESPRIT algorithm can achieve a similar performance as 2-D DFT ESPRIT without these phase jump errors.

Traditional problems in computational geometry involve aspects that are both discrete and continuous. One such example is nearest-neighbor searching, where the input is discrete, but the result depends on distances, which vary continuously. In many real-world applications of geometric data structures, it is assumed that query results are continuous, free of jump discontinuities. This is at odds with many modern data structures in computational geometry, which employ approximations to achieve efficiency, but these approximations often suffer from discontinuities. In this paper, we present a general method for transforming an approximate but discontinuous data structure into one that produces a smooth approximation, while matching the asymptotic space efficiencies of the original. We achieve this by adapting an approach called the partition-of-unity method, which smoothly blends multiple local approximations into a single smooth global approximation. We illustrate the use of this technique in a specific application of approximating the distance to the boundary of a convex polytope in $\mathbb{R}^d$ from any point in its interior. We begin by developing a novel data structure that efficiently computes an absolute $\varepsilon$-approximation to this query in time $O(\log (1/\varepsilon))$ using $O(1/\varepsilon^{d/2})$ storage space. Then, we proceed to apply the proposed partition-of-unity blending to guarantee the smoothness of the approximate distance field, establishing optimal asymptotic bounds on the norms of its gradient and Hessian.

Approximate nearest-neighbor search is a fundamental algorithmic problem that continues to inspire study due its essential role in numerous contexts. In contrast to most prior work, which has focused on point sets, we consider nearest-neighbor queries against a set of line segments in ℝ^d, for constant dimension d. Given a set S of n disjoint line segments in ℝ^d and an error parameter ε > 0, the objective is to build a data structure such that for any query point q, it is possible to return a line segment whose Euclidean distance from q is at most (1+ε) times the distance from q to its nearest line segment. We present a data structure for this problem with storage O((n²/ε^d) log (Δ/ε)) and query time O(log (max(n,Δ)/ε)), where Δ is the spread of the set of segments S. Our approach is based on a covering of space by anisotropic elements, which align themselves according to the orientations of nearby segments.

Transfer learning facilitates the training of task-specific classifiers using pre-trained models as feature extractors. We present a family of transferable adversarial attacks against such classifiers, generated without access to the classification head; we call these headless attacks. We first demonstrate successful transfer attacks against a victim network using only its feature extractor. This motivates the introduction of a label-blind adversarial attack. This transfer attack method does not require any information about the class-label space of the victim. Our attack lowers the accuracy of a ResNet18 trained on CIFAR10 by over 40%.

Despite the vulnerability of object detectors to adversarial attacks, very few defenses are known to date. While adversarial training can improve the empirical robustness of image classifiers, a direct extension to object detection is very expensive. This work is motivated by recent progress on certified classification by randomized smoothing. We start by presenting a reduction from object detection to a regression problem. Then, to enable certified regression, where standard mean smoothing fails, we propose median smoothing, which is of independent interest. We obtain the first model-agnostic, training-free, and certified defense for object detection against $\ell_2$-bounded attacks.

Polyhedral meshes are increasingly becoming an attractive option with particular advantages over traditional meshes for certain applications. What has been missing is a robust polyhedral meshing algorithm that can handle broad classes of domains exhibiting arbitrary curved boundaries and sharp features. In addition, the power of primal-dual mesh pairs, exemplified by Voronoi-Delaunay meshes, has been recognized as an important ingredient in numerous formulations. The VoroCrust algorithm is the first provably correct algorithm for conforming Voronoi meshing for non-convex and possibly non-manifold domains with guarantees on the quality of both surface and volume elements. A robust refinement process estimates a suitable sizing field that enables the careful placement of Voronoi seeds across the surface circumventing the need for clipping and avoiding its many drawbacks. The algorithm has the flexibility of filling the interior by either structured or random samples, while all sharp features are preserved in the output mesh. We demonstrate the capabilities of the algorithm on a variety of models and compare against state-of-the-art polyhedral meshing methods based on clipped Voronoi cells establishing the clear advantage of VoroCrust output.

We present a new approach to ε-approximate nearest-neighbor queries in fixed dimension under a variety of non-Euclidean distances. We consider two families of distance functions: (a) convex scaling distance functions including the Mahalanobis distance, the Minkowski metric and multiplicative weights, and (b) Bregman divergences including the Kullback-Leibler divergence and the Itakura-Saito distance. As the fastest known data structures rely on the lifting transformation, their application is limited to the Euclidean metric, and alternative approaches for other distance functions are much less efficient. We circumvent the reliance on the lifting transformation by a careful application of convexification, which appears to be relatively new to computational geometry. We are given n points in ℝd, each a site possibly defining its own distance function. Under mild assumptions on the growth rates of these functions, the proposed data structures answer queries in logarithmic time using O(n log(1/ε)/εd/2) space, which nearly matches the best known results for the Euclidean metric.

Low-cost mini-drones with advanced sensing and maneuverability enable a new class of intelligent sensing systems. To achieve the full potential of such drones, it is necessary to develop new enhanced formulations of both common and emerging sensing scenarios. Namely, several fundamental challenges in visual sensing are yet to be solved including (1) fitting sizable targets in camera frames; (2) positioning cameras at effective viewpoints matching target poses; and (3) accounting for occlusion by elements in the environment, including other targets. In this article, we introduce Argus, an autonomous system that utilizes drones to collect target information incrementally through a two-tier architecture. To tackle the stated challenges, Argus employs a novel geometric model that captures both target shapes and coverage constraints. Recognizing drones as the scarcest resource, Argus aims to minimize the number of drones required to cover a set of targets. We prove this problem is NP-hard, and even hard to approximate, before deriving a best-possible approximation algorithm along with a competitive sampling heuristic which runs up to 100× faster according to large-scale simulations. To test Argus in action, we demonstrate and analyze its performance on a prototype implementation. Finally, we present a number of extensions to accommodate more application requirements and highlight some open problems.

times the local feature size centered at each sample. The corners of the union of these balls on both sides of the surface are the Voronoi sites and the interface of their cells is a watertight surface reconstruction embedded in the dual shape of the union of balls. With the surface protected, the enclosed volume can be further decomposed by generating more sites inside it. Compared to clipping-based algorithms, VoroCrust cells are full Voronoi cells, with convexity and fatness guarantees. Compared to the power crust algorithm, VoroCrust cells are not filtered, are unweighted, and offer greater flexibility in meshing the enclosed volume by either structured or randomly genenerated samples.

Over the past decade, polyhedral meshing has been gaining popularity as a better alternative to tetrahedral meshing in certain applications. Within the class of polyhedral elements, Voronoi cells are particularly attractive thanks to their special geometric structure. What has been missing so far is a Voronoi mesher that is sufficiently robust to run automatically on complex models. In this video, we illustrate the main ideas behind the VoroCrust algorithm, highlighting both the theoretical guarantees and the practical challenges imposed by realistic inputs.

Low-cost mini-drones with advanced sensing and maneuverability enable a new class of intelligent visual sensing systems. This potential motivated several research efforts to employ drones as standalone surveillance systems or to assist legacy deployments. However, several fundamental challenges remain unsolved including: 1) Adequate coverage of sizable targets; 2) Target orientation that render coverage effective only from certain directions; 3) Occlusion by elements in the environment, including other targets.

Abstract In many geometry processing applications, it is required to improve an initial mesh in terms of multiple quality objectives. Despite the availability of several mesh generation algorithms with provable guarantees, such generated meshes may only satisfy a subset of the objectives. The conflicting nature of such objectives makes it challenging to establish similar guarantees for each combination, e.g., angle bounds and vertex count. In this paper, we describe a versatile strategy for mesh improvement by interpreting quality objectives as spatial constraints on resampling and develop a toolbox of local operators to improve the mesh while preserving desirable properties. Our strategy judiciously combines smoothing and transformation techniques allowing increased flexibility to practically achieve multiple objectives simultaneously. We apply our strategy to both planar and surface meshes demonstrating how to simplify Delaunay meshes while preserving element quality, eliminate all obtuse angles in a complex mesh, and maximize the shortest edge length in a Voronoi tessellation far better than the state‐of‐the‐art.

The NewsStand system enables the use of a map query interface to retrieve news articles associated with the principal locations that they mention collected as a result of monitoring the output of over 10,000 RSS news feeds, made available within minutes of publication. NewsStand has been enhanced to allow using the map query interface to access other information associated with the articles such as photos and videos, as well as names of people and diseases mentioned in these articles. Here we report on our efforts to enhance NewsStand to display the names of brands and to the articles mentioning them. The challenges in identifying interesting brand mentions are discussed.

The problem of transmit beamforming for single-group multicasting is considered, where the objective is to transmit common information to a (large) number of users. The transmitter is assumed to have accurate downlink channel state information (CSI) for all users, and the objective is to design the beamformer weights to minimize the total transmitted power subject to meeting the quality-of-service (QoS) constraints of all users. This is an NP-hard problem that has recently drawn considerable interest (e.g., in the context of UMTS-LTE / E-MBMS). Several channel orthogonalization-based methods are proposed to solve this problem in an approximate way. Our techniques are shown to offer an improved performance-to-complexity tradeoff as compared to the original semidefinite relaxation (SDR) based multicasting technique.