Active Cloaking for an Object in General Rectilinear Motion

The main objective of this study was to investigate how ultra-fast CT scanning influences cardiac pulsation artefacts and image quality in low dose chest CT, clinically and in a non-moving 3D resolution phantom. 30 patients underwent two non-contrast chest CT examinations within 6months using one ultra-fast and one standard speed protocol. Three radiologists scored general image quality and motion artefacts (ascending aorta, superior aortic recess, left ventricle, lung tissue) using a 4-point Likert scale. Lung parenchyma, pulmonary fissures, segmental bronchi, pleuromediastinal border, carina and pulmonary vessels were compared between ultra-fast and standard speed scans. Volumetric resolution was measured with an image quality phantom containing a 3D resolution wave module to assess the non-motion resolution on both protocols. With ultra-fast scans 59-70% of the investigated areas in 30 patients were scored as having no artefacts versus 10-29% of the standard speed scans (p<0,001). General image quality was scored excellent or good in 94% of the cases with ultra-fast scans and in 59% of the cases with standard speed scans (p<0,001). Comparing anatomical structures, ultra-fast scans were considered equal or better than standard speed scans in 78%/90%/93%/92%/91%/74% of the examinations (lung parenchyma/pulmonary fissures/segmental bronchi/pleuromediastinal border/carina/pulmonary vessels). Phantom scans showed a slight reduction in volumetric resolution MTFxyz, and no change in in-plane resolution, MTFxy. In this study, ultra-fast scans reduce motion artefacts and increase image quality in low dose non-contrast chest CT when compared to standard speed scans, despite a possible slight degradation in resolution in non-moving objects.

Authoring high-quality character animation is essential in multimedia production, especially for pre-rendered formats such as films and TV series. These workflows are often iterative, requiring frequent adjustments and immediate visual feedback, and typically involve editing existing motion, such as mocap data. We present Neural Motion Path (NMP), a deep learning-based system designed to support this authoring process by enabling full-body motion editing through joint-level motion path manipulation. While joint rotations are essential for expressive human motion, explicitly specifying them imposes a significant burden on users; NMP addresses this challenge by enabling intuitive position-only motion path editing while implicitly inferring plausible rotation trajectories. To address the inherent tension between motion plausibility and precise constraint satisfaction, NMP explicitly decouples motion synthesis from constraint enforcement within an autoregressive framework. NMP generates realistic, context-aware motion without fine-tuning every path detail, making it accessible to novices while supporting the demands of detailed animation authoring. It combines a motion generator with a novel RotationNet for inferring joint rotation, and a constraint imposer that enforces end-effector constraints via an Explicit-Weight Sparse Expert Model (EW-SEM). The system supports terrain adaptation and authoring operations like Concatenate, Insert, and Mix. Implemented as a Blender add-on, NMP supports real-time playback and interactive workflows. A user study with novice users shows that NMP improves satisfaction, efficiency, and perceived motion quality compared to the conventional layered keyframing approach, highlighting its potential as an accessible and effective authoring tool.

Recent advances in generative models have yielded impressive progress on motion in-betweening, allowing for more complex, varied, and realistic motion transitions. However, recent methods still exhibit noticeable limitations in preserving keyframe information and ensuring motion continuity. In this paper, we propose a novel pipeline and sampling optimization strategy for latent diffusion models (LDM) based on motion implicit neural representations (INR). By establishing a mapping between INR and sparse spatial or temporal information within latent diffusion, our model can sample the INR parameters from extremely sparse and ambiguous keyframe data and reconstruct plausible and smooth motions from the manifold. Our experiments demonstrate the superior performance of our model, which significantly improves motion generation quality in scenarios with few keyframes while ensuring both keyframe accuracy and diversity of in-between motions.

This work aims at a challenging task: human action-reaction synthesis, i.e., generating human reactions conditioned on the action sequence of another person. Currently, autoregressive modeling approaches with vector quantization (VQ) have achieved remarkable performance in motion generation tasks. However, VQ has inherent disadvantages, including quantization information loss, low codebook utilization, etc. In addition, while dividing the body into separate units can be beneficial, the computational complexity needs to be considered. Also, the importance of mutual perception among units is often neglected. In this work, we propose MARRS, a novel framework designed to generate coordinated and fine-grained reaction motions using continuous representations. Initially, we present the Unit-distinguished Motion Variational AutoEncoder (UD-VAE), which segments the entire body into distinct body and hand units, encoding each independently. Subsequently, we propose Action-Conditioned Fusion (ACF), which involves randomly masking a subset of reactive tokens and extracting specific information about the body and hands from the active tokens. Furthermore, we introduce Mutual Unit Modulation (MUM) to facilitate interaction between body and hand units by using the information from one unit to adaptively modulate the other. Finally, for the diffusion model, we employ a compact MLP as a noise predictor for each distinct body unit and incorporate the diffusion loss to model the probability distribution of each token. Both quantitative and qualitative results demonstrate that our method achieves superior performance.

Walking is fundamental to humans and animals, yet the neural principles underlying movement generation remain unclear. In particular, the relationship between neuronal cell types, networks, and functions has been difficult to establish. Here, we propose that the spatial organization of the spinal cord governs network-driven locomotor rhythms and patterns. An asymmetric "Mexican hat" connectivity - local excitation with longer-range inhibition and a longitudinal skew - accounts for proper motor dynamics, while segregation of cell types in the transversal plane allows descending fibers to find appropriate targets and control network dynamics. We extract these principles via a model of the mouse spinal cord, where synaptic connections are sampled probabilistically from cell-specific projection patterns derived from single-cell RNA sequencing and spatial transcriptomics. Essential aspects of locomotion are induced and controlled without extensive parameter optimization. We additionally predict propagating activity bumps during rhythmic movement. This work reveals universal spatial principles linking cell types, connectivity, and behavior across species.

Low-frame-rate flat-panel detector and gantry design limit Cone-beam computed tomography (CBCT) scanning speed, increasing susceptibility to patient motion during prolonged scanning. Rigid motion invalidates the predetermined imaging geometry, leading to degraded image quality. To ensure diagnostic reliability and support downstream tasks, accurate motion estimation and subsequent motion-compensated reconstruction are essential. This work presents a general rigid motion correction framework for CBCT using differentiable 3D-2D registration. We introduce a Lie group manifold-constrained motion modeling approach that conforms to the definition of rigid transformation and analytically derive the gradient of the CBCT forward projection operator with respect to rigid motion parameters. Furthermore, we identify the spatial deviation relationship between motion-corrupted and ground truth images, proposing a motion estimation constraint to enhance convergence. Experimental comparisons with state-of-the-art methods demonstrate that our method achieves superior motion estimation and image restoration. It effectively recovers anatomical details and substantially reduces rigid motion artifacts, even under severe motion conditions.

In the areas of robotics, sports science, virtual reality and healthcare, there has been much interest placed on human motion generation and analysis. With precise human motion modeling, HCI may enable the computer systems to comprehend, categorize, and predict these actions. Deep learning and optical motion capture technologies have opened a window to the extraction of rich spatial-temporal features from human actions for improvement in realistic motion generation and real-time analysis. A common problem faced by modern motion analysis systems is to account for the complexity in joint relations and to maintain temporal continuity, resulting in jerky and unrealistic motions during output. Additionally, some of the models are not very generalizable, particularly because they are sensitive to speed and style changes. Most conventional techniques of anomaly detection rely on manually derived thresholds, limiting their adaptability in dynamic real-world situations. This research aims to combine optical motion capture data with spatial-temporal graph convolutional networks (ST-GCNs) to overcome these limitations and establish a unified framework for motion generation and analysis. The aim is to provide reliable, scalable, realistic motion synthesis, anomaly detection, and accurate motion recognition solutions. The work is also geared towards improving model accuracy and adaptability through an efficient preprocessing and data augmentation technique. The ST-GCN spatiotemporal method for human motion modeling uses graph representations (nodes and edges) of joint movements, while the data preprocessing stages are cleaning, skeleton normalization, and data augmentation through mirroring and time scaling. The subsequent step is to build a model using a graph-based deep learning technique for motion generation, anomaly detection, and action recognition tasks. The ST-GCN model showed a high level of realism, with KL divergence values of 0.35 for “Jump” motion synthesis, while 98% action recognition accuracy was achieved for tasks like “Run” and “Punch.” Outlier anomalies detection through a threshold of 0.7 proved to be sufficient to ‘catch’ any such instances, showcasing the capabilities of the model in both analysis and generation. These results bear testimony to the effectiveness of the framework toward resolving contemporary constraining issues in motion analysis and generation. Work Related Contributions: The unified ST-GCN architecture for combined motion analysis and generation, robust preprocessing, and augmentation for improved generalization and incorporation of the graph-based anatomical constraint for preserving the kinematic constraints. It further integrates with a generative ST-GCN extension for a realistic and smooth synthesis and with adaptive anomaly detection, replacing fixed thresholds with learned spatio-temporal patterns, therefore improving novelty, usability, and scalability.

Detailed perceptual discrimination is suggested to be inferior in peripheral vision, with sensitivity declining continuously as retinal eccentricity increases. Yet most ecologically relevant events, including imminent threats, are first signaled in the periphery, raising the possibility that peripheral capabilities for threat processing have been underestimated. Here, we tested whether peripheral vision preserves discrimination sensitivity to the trajectories of looming stimuli-visual patterns simulating collision-which are known to engage rapid subcortical visual pathways with large receptive fields. Across five psychophysical experiments in desktop and virtual reality (VR) settings, we consistently found that discrimination sensitivity for looming stimuli remained stable from parafoveal to near-peripheral eccentricities (2° to 14.5°) and declined only at the far periphery (30°). In contrast, physically matched receding stimuli showed typical eccentricity-dependent decline, ruling out general motion sensitivity as the source of preserved performance. Pupil responses further supported the affective salience of looming stimuli. These findings suggest that peripheral vision may be selectively optimized for detecting biologically relevant threats via subcortical mechanisms, challenging the traditional fovea-centered view of visual discrimination.

HLM-MOCAP: A motion capture dataset of context-dependent human arm motion for human-like robot motion generation

Modern methods of aquifer hydraulic testing estimate subsurface properties by creating frequency-dependent variations in groundwater well levels. These methods are variously known as periodic, harmonic, or oscillatory hydraulic testing, of which sinusoidal testing is a specific case. Periodic testing provides several potential advantages over traditional methods, including larger signal-to-noise ratios, larger distances over which hydraulic disturbances propagate, and the ability to undertake zero net water extraction. One of three approaches are used to induce groundwater pressure fluctuations: (1) extraction/reinjection of water using motorized pumps, (2) pressurization/depressurization using compressed air, and (3) transient displacement of the water column by slug testing. The latter was the focus of the present study; specifically, how to improve sinusoidal slug testing methods by ensuring accurate generation of sinusoidal variations in well water levels. Two previously published sinusoidal testing designs were evaluated in terms of the ratio of effective transfer link length, , to effective flywheel radius, . The first published design featured ratio values ranging from 13 to 9, which corresponded to maximum discrepancies between intended and actual slug movement of 7% to 10%, respectively. The second design featured ratio values ranging from 12 to 2, which corresponded to maximum discrepancies of 8% to 29%, respectively. These analyses suggest that methods featuring a rotating drive coupled to an effective transfer link are suboptimal. Instead, designs featuring either modified flywheel apparatus or winches driven by digitally controlled stepper motors can minimize the potential for discrepancies between intended and actual slug movement. The accurate generation of sinusoidal slug movement will minimize uncertainties associated with hydraulic properties inversely estimated from observations acquired during sinusoidal slug testing.

ABSTRACT Music‐driven motion generation has attracted increasing attention, yet conducting remains underexplored despite its central role in musical communication. We address this gap with a focus on choral conducting. We introduce Maestro3D, a large‐scale 3D dataset of professional performances, and propose a beat‐aware diffusion framework enhanced by a novel phase‐based beat representation that explicitly encodes rhythmic structure. Both quantitative and qualitative evaluations show that our approach achieves superior accuracy, realism, and synchronization compared to existing methods.

Three-Sensor Inertial Poser (3SIP): Full-body motion generation using only three inertial measurement units

ABSTRACT Real‐time character animation requires generating natural motions while preserving diverse walking styles under user control. Phase‐based representations are commonly employed in motion generation frameworks to control periodic motions such as walking; however, existing approaches face a trade‐off between preserving fine‐grained local periodic details and maintaining coherent whole‐body motion. Methods focusing on local periodic features often insufficiently represent other joints, while global phase representations tend to smooth out stylistic details, leading to style homogenization. This paper proposes an end‐effector‐aware phase manifold learning framework that balances local stylistic features and global motion consistency. The proposed method employs a two‐stage training strategy that first learns local periodic characteristics of end‐effectors and then integrates full‐body periodicity while fixing the learned local representations. In addition, we introduce an angular velocity‐based phase representation, which more directly captures the rotational characteristics of walking motions than linear velocity. Experimental results demonstrate improved style preservation for gesture‐dominant motions while generating stable whole‐body walking motions.

Holistic Co-Speech Motion Generation via Cross-Gated Attention and Cross-Limb Interaction

Talking head generation is increasingly important in virtual reality (VR), especially for social scenarios involving multi-turn conversation. Existing approaches face notable limitations: mesh-based 3D methods can model dual-person dialogue but lack realistic textures, large-model-based 2D methods produce natural appearances but incur prohibitive computational costs. Recently, 3D Gaussian Splatting (3DGS)-based methods achieve efficient and realistic rendering but remain speaker-only and ignore social relationships. We introduce RSATalker, the first framework that leverages 3DGS for realistic and socially-aware talking head generation, with support for multi-turn conversation. Our method first drives mesh-based 3D facial motion from speech, then binds 3D Gaussians to mesh facets to render high-fidelity 2D avatar videos. To capture interpersonal dynamics, we propose a socially-aware module that encodes social relationships, including blood and non-blood as well as equal and unequal, into high-level embeddings through a learnable query mechanism. We design a three-stage training paradigm and construct the RSATalker dataset with speech-mesh-image triplets annotated with social relationships. Our method supports applications such as VR telepresence, social VR, and embodied conversational agents. The socially-aware conditioning can also be extended to other human motion generation tasks. Extensive experiments demonstrate that RSATalker achieves state-of-the-art performance in both realism and social awareness. The code and dataset will be released.

Online multi-object tracking (MOT) is typically implemented as tracking-by-detection: a motion prior propagates each track to the current frame, and appearance cues drive data association. Despite strong detectors and association heuristics, two practical failure modes persist in crowded, interaction-heavy videos: (i) deterministic, unimodal propagation over-commits under non-linear motion and short-term ambiguity, causing overlap-based gating to discard correct matches; and (ii) appearance embeddings drift over time, so stale historical features can dominate similarity scores after long gaps. This paper proposes DiffuTrack, a generative online MOT framework that addresses both issues while retaining the standard predict-associate-update loop. The Motion Diffusion Module (MDM) replaces point motion propagation with a conditional diffusion generator over normalized bounding-box states, producing distributional hypotheses via accelerated DDIM sampling. To stabilize identity association, Time-Aware Prototype Contrastive Learning (TPCL) maintains temporally decayed identity prototypes and trains embeddings with a time-weighted contrastive objective that down-weights stale positives. Experiments on MOT17/MOT20 and DanceTrack under a shared-detection pedestrian-tracking protocol show consistent gains in association-centric metrics, with the largest improvements concentrated on highly non-linear trajectories and occlusion-heavy sequences. Diagnostic uncertainty-coverage analyses further indicate that diffusion-based hypotheses retain a wider region of support than linear-Gaussian propagation, supporting probabilistic motion generation as a practical alternative to deterministic tracking priors in the standard online MOT loop.

Recurrent neural networks (RNNs) with predefined-time convergence capabilities are among the most powerful solvers for time-varying zero-finding problems (TVZFPs). However, a comprehensive design framework for such neural networks has not yet been well established. To address this gap, this article presents a comprehensive framework for generating a series of adaptive arbitrarily predefined-time convergent RNNs (A-APTC-RNNs). Compared with most existing RNNs, the A-APTC-RNNs generated using the proposed comprehensive framework exhibit the following distinctive features: 1)owing to a novel piecewise evolution formula, their convergence time can be arbitrarily predefined; 2)owing to a proportional-integral-derivative regulatory mechanism, they achieve lower steady-state residual errors after convergence; and 3)owing to a novel adaptive parameter initialization scheme, they are able to automatically determine their own model parameters. Theoretical analysis rigorously demonstrates the stability and arbitrarily predefined-time convergence (APTC) capability of the A-APTC-RNNs. Various experiments (i.e., numerical simulations, alternating-current estimation, chaotic synchronization of Chua's circuit, and motion generation for dual-arm robots) demonstrate the state-of-the-art convergence performance of the A-APTC-RNNs generated by the proposed comprehensive framework.

Abstract In spherical piezoelectric motors (SPM) attitude control ensures precise rotor orientation, thereby improving reliability, minimizing errors, and optimizing efficiency. The existing studies about the SPM system have mainly focused on multi-axial torque output while neglecting the techniques for controlling the spherical motor’s rotor attitude. Most of the SPM systems are limited to their range of orientations and can achieve rotations only along the x , y , and z directions. Except for a few systems capable of providing multi-axial moments, such systems are typically confined to a single 2D-plane or their axial movement ranges are not well-defined and mostly rely on open-loop control systems without model-based control methods. To tackle these issues, in this article, a feedback-control framework for rotor attitude control has been proposed. The rotor’s driving torques are specifically configured to attain omnidirectional capability. This is achieved through a linear combination of six elementary torque orientations, achieved by adjusting the positions of the contact points between the stators and the rotor. This study adopts a quaternion-based dynamic framework to effectively eliminate singularities in the manipulation of omnidirectional rotors and systematically derives independent control commands for each stator using a computed torque control approach, making it suitable for motor systems that function across all axial orientations. The study implements the developed rotor attitude control strategy in general motion control applications, which are broadly classified into three operational categories: positioning, spin, and hybrid modes. A priority level-based switching algorithm is developed to facilitate torque trajectory generation and reduce the dynamic effects caused by switching rotor contact positions. Finally, a simulation is conducted to validate the proposed rotor attitude control algorithm. The results confirm that the introduced approach delivers consistent and flexible performance suitable for a wide spectrum of motion control scenarios. Simulation analyses further demonstrate that the integration of the switching algorithm enables the rotor to consistently achieve its target orientations throughout different spatial regions.

&lt;p&gt;Senior adults are a growing population in America and face increased feelings of social isolation, a problem compounded by the COVID-19 pandemic. The positive physiological impacts of dance fitness classes for seniors are well documented, but less research exists on the impacts of creative dance on well-being and social connection. This study addresses this gap in knowledge by focusing on the expressive potential and social impact of senior adult engagement in creative dance. Over the course of a three-week dance and storytelling workshop, senior adult participants experienced an outlet for creative expression and imagination. The methods include oral history interviews, a group interview, facilitator field notes, oral storytelling, movement generation from language, creative writing exercises and somatic guided movement experiences. The study is situated within participatory action research (PAR), with regular dialogue and participant feedback guiding the workshop. In the context of community engaged dance, this research reveals the expressive potential that comes from bridging language and movement. I will show how utilizing both a “language-first” approach and a “movement-first” approach engaged the seniors’ imaginations in what Vygotsky terms “combinatorial creativity.” These approaches contributed to positive social connection outcomes and a sense of self-discovery among the senior adult participants and suggest the importance of imagination for creative expression in future community-based dance engagement.&lt;/p&gt;