Search

Text logo design heavily relies on the creativity and expertise of professional designers, in which arranging element layouts is one of the most important procedures. However, few attention has been paid to this specific task which needs to take precise textural details and user constraints into consideration, but only on the broader tasks such as document/poster layout generation. In this paper, we propose a VLM-based framework that generates content-aware text logo layouts by integrating multi-modal inputs with user constraints, supporting a more flexible and stable layout design in real-world applications. We introduce two model techniques to reduce the computation for processing multiple glyph images simultaneously, while does not face performance degradation. To support instruction-tuning of out model, we construct two extensive text logo datasets, which are 5x more larger than the existing public dataset. Except for the geometric annotations (e.g. text masks and character recognition), we also compliment with comprehensive layout descriptions in natural language format, for more effective training to have reasoning ability when dealing with complex layouts and custom user constraints. Experimental studies demonstrate the effectiveness of our proposed model and datasets, when comparing with previous methods in various benchmarks to evaluate geometric aesthetics and human preferences. The code and datasets will be publicly available.

MetaDesigner revolutionizes artistic typography synthesis by leveraging the strengths of Large Language Models (LLMs) to drive a design paradigm centered around user engagement. At the core of this framework lies a multi-agent system comprising the Pipeline, Glyph, and Texture agents, which collectively enable the creation of customized WordArt, ranging from semantic enhancements to the imposition of complex textures. MetaDesigner incorporates a comprehensive feedback mechanism that harnesses insights from multimodal models and user evaluations to refine and enhance the design process iteratively. Through this feedback loop, the system adeptly tunes hyperparameters to align with user-defined stylistic and thematic preferences, generating WordArt that not only meets but exceeds user expectations of visual appeal and contextual relevance. Empirical validations highlight MetaDesigner's capability to effectively serve diverse WordArt applications, consistently producing aesthetically appealing and context-sensitive results., Comment: 18 pages, 16 figures, Project: https://modelscope.cn/studios/WordArt/WordArt

A vortex, a circulating flow around a void, is one of the basic topological phenomena in nature. Here we show that vortices generally emerge in spin wave travelling upon topologically nontrivial magnetic texture, due to the transverse precession of spin wave about the background magnetization. The winding number of each spin wave vortex is equivalent to sign of the local topological density of magnetic texture at the vortex core, and all winding numbers add up as twice the topological number of the magnetic texture. Based on the charts of spin wave vortices, the magnetization profile of the magnetic texture is reversely constructed, and a universal relation for the magnon topological Hall angle is theoretically proposed and numerically confirmed in vast types of magnetic textures. The simple connection between dynamic and static magnetizations, promotes spin wave vortex as a powerful tool to reveal the topology of the underlying magnetic texture., Comment: 7 pages,4 figures. Comments are welcome

Multimodal Large Language Model (MLLMs) leverages Large Language Models as a cognitive framework for diverse visual-language tasks. Recent efforts have been made to equip MLLMs with visual perceiving and grounding capabilities. However, there still remains a gap in providing fine-grained pixel-level perceptions and extending interactions beyond text-specific inputs. In this work, we propose {\bf{AnyRef}}, a general MLLM model that can generate pixel-wise object perceptions and natural language descriptions from multi-modality references, such as texts, boxes, images, or audio. This innovation empowers users with greater flexibility to engage with the model beyond textual and regional prompts, without modality-specific designs. Through our proposed refocusing mechanism, the generated grounding output is guided to better focus on the referenced object, implicitly incorporating additional pixel-level supervision. This simple modification utilizes attention scores generated during the inference of LLM, eliminating the need for extra computations while exhibiting performance enhancements in both grounding masks and referring expressions. With only publicly available training data, our model achieves state-of-the-art results across multiple benchmarks, including diverse modality referring segmentation and region-level referring expression generation., Comment: CVPR 2024

Depth-aware panoptic segmentation is an emerging topic in computer vision which combines semantic and geometric understanding for more robust scene interpretation. Recent works pursue unified frameworks to tackle this challenge but mostly still treat it as two individual learning tasks, which limits their potential for exploring cross-domain information. We propose a deeply unified framework for depth-aware panoptic segmentation, which performs joint segmentation and depth estimation both in a per-segment manner with identical object queries. To narrow the gap between the two tasks, we further design a geometric query enhancement method, which is able to integrate scene geometry into object queries using latent representations. In addition, we propose a bi-directional guidance learning approach to facilitate cross-task feature learning by taking advantage of their mutual relations. Our method sets the new state of the art for depth-aware panoptic segmentation on both Cityscapes-DVPS and SemKITTI-DVPS datasets. Moreover, our guidance learning approach is shown to deliver performance improvement even under incomplete supervision labels., Comment: to be published in ICCV 2023

Visual object tracking is a fundamental video task in computer vision. Recently, the notably increasing power of perception algorithms allows the unification of single/multiobject and box/mask-based tracking. Among them, the Segment Anything Model (SAM) attracts much attention. In this report, we propose HQTrack, a framework for High Quality Tracking anything in videos. HQTrack mainly consists of a video multi-object segmenter (VMOS) and a mask refiner (MR). Given the object to be tracked in the initial frame of a video, VMOS propagates the object masks to the current frame. The mask results at this stage are not accurate enough since VMOS is trained on several closeset video object segmentation (VOS) datasets, which has limited ability to generalize to complex and corner scenes. To further improve the quality of tracking masks, a pretrained MR model is employed to refine the tracking results. As a compelling testament to the effectiveness of our paradigm, without employing any tricks such as test-time data augmentations and model ensemble, HQTrack ranks the 2nd place in the Visual Object Tracking and Segmentation (VOTS2023) challenge. Code and models are available at https://github.com/jiawen-zhu/HQTrack., Comment: Technical Report

We show that spin wave acquires a polarization-dependent geometric phase along a cyclic trajectory of non-coplanar magnetizations in antiferromagnets. Specifically, we demonstrate that a cyclic set of 90 degree antiferromagnetic domain walls simultaneously introduce geometric and dynamic phases to spin wave, and thus leads to asymmetric magnitude of overall phase for left-/right-circular components. Based on the polarization-dependent phase, we propose theoretically and confirm by micromagnetic simulations that, a Mach-Zehner interferometer with cyclic 90 degree domain walls in one arm and homogenous domain in the other arm, naturally acts as a spin wave circular polarizer. Moreover, the circular polarizer has intrinsic nonreciprocity, which filters opposite polarization in opposite propagation direction., Comment: 6 pages, 3 figures

Magnetic skyrmions widely exist in a diverse range of magnetic systems, including chiral magnets with a non-centrosymmetric structure characterized by Dzyaloshinkii-Moriya interaction~(DMI). In this study, we propose a generalized semi-implicit backward differentiation formula projection method, enabling the simulations of the Landau-Lifshitz~(LL) equation in chiral magnets in a typical time step-size of $1$ ps, markedly exceeding the limit subjected by existing numerical methods of typically $0.1$ ps. Using micromagnetics simulations, we show that the LL equation with DMI reveals an intriguing dynamic instability in magnetization configurations as the damping varies. Both the isolated skyrmionium and skyrmionium clusters can be consequently produced using a simple initialization strategy and a specific damping parameter. Assisted by the string method, the transition path between skyrmion and skyrmionium, along with the escape of a skyrmion from the skyrmion clusters, are then thoroughly examined. The numerical methods developed in this work not only provide a reliable paradigm to investigate the skyrmion-based textures and their transition paths, but also facilitate the understandings for magnetization dynamics in complex magnetic systems.

The interplay between spin wave and magnetic texture represents the information exchange between the fast and slow dynamical parts of magnetic systems. Here we formulate a set of magnonic torques acting on background magnetic texture, by extracting time-invariant information from the fast precessing spin waves. Under the frame of magnonic torques, we use theoretical formulations and micromagnetic simulations to investigate the spin wave driven domain wall motion in two typical symmetry-breaking situations: the rotational symmetry broken by the Dzyaloshinkii-Moriya interaction, and the translational symmetry broken by magnetic damping. The torque-based microscopic analyses provide compact yet quantitative tools to reinterpret the magnetic texture dynamics induced by spin wave, beyond the conventional framework of global momentum conservation., Comment: 12 pages, 4 figures

Streaming perception is a critical task in autonomous driving that requires balancing the latency and accuracy of the autopilot system. However, current methods for streaming perception are limited as they only rely on the current and adjacent two frames to learn movement patterns. This restricts their ability to model complex scenes, often resulting in poor detection results. To address this limitation, we propose LongShortNet, a novel dual-path network that captures long-term temporal motion and integrates it with short-term spatial semantics for real-time perception. LongShortNet is notable as it is the first work to extend long-term temporal modeling to streaming perception, enabling spatiotemporal feature fusion. We evaluate LongShortNet on the challenging Argoverse-HD dataset and demonstrate that it outperforms existing state-of-the-art methods with almost no additional computational cost., Comment: Accepted at ICASSP 2023, source code is at https://github.com/zhiqic/LongShortNet

Existing Visual Object Tracking (VOT) only takes the target area in the first frame as a template. This causes tracking to inevitably fail in fast-changing and crowded scenes, as it cannot account for changes in object appearance between frames. To this end, we revamped the tracking framework with Progressive Context Encoding Transformer Tracker (ProContEXT), which coherently exploits spatial and temporal contexts to predict object motion trajectories. Specifically, ProContEXT leverages a context-aware self-attention module to encode the spatial and temporal context, refining and updating the multi-scale static and dynamic templates to progressively perform accurately tracking. It explores the complementary between spatial and temporal context, raising a new pathway to multi-context modeling for transformer-based trackers. In addition, ProContEXT revised the token pruning technique to reduce computational complexity. Extensive experiments on popular benchmark datasets such as GOT-10k and TrackingNet demonstrate that the proposed ProContEXT achieves state-of-the-art performance., Comment: Accepted at ICASSP 2023, source code is at https://github.com/zhiqic/ProContEXT

Polarization, the precession direction with respect to the background magnetization, is an intrinsic degree of freedom of spin wave. Introducing symmetry breaking mechanisms lies in the heart of lifting the degeneracy between polarized spin wave modes, and is essential in constructing polarization-based magnonic devices. Here we show that polarized spin waves can be naturally harnessed in a curved antiferromagnetic wire via tuning its curvature and torsion. Specifically, we investigate evolution of polarized spin wave in a spin wave rotator and a spin wave interferometer based on magnetic circular helices, and correlate these curvilinear effects to the Berry phase accumulated along wires., Comment: 6 pages, 3 figures

Recent theoretical and experimental advances show that the inertia of magnetization emerges at sub-picoseconds and contributes to the ultrafast magnetization dynamics which cannot be captured intrinsically by the LLG equation. Therefore, as a generalization, the inertial Landau-Lifshitz-Gilbert (iLLG) equation is proposed to model the ultrafast magnetization dynamics. Mathematically, the LLG equation is a nonlinear system of parabolic type with (possible) degeneracy. However, the iLLG equation is a nonlinear system of mixed hyperbolic-parabolic type with degeneracy, and exhibits more complicated structures. It behaves like a hyperbolic system at the sub-picosecond scale while behaves like a parabolic system at larger timescales. Such hybrid behaviors impose additional difficulties on designing numerical methods for the iLLG equation. In this work, we propose a second-order semi-implicit scheme to solve the iLLG equation. The second temporal derivative of magnetization is approximated by the standard centered difference scheme and the first derivative is approximated by the midpoint scheme involving three time steps. The nonlinear terms are treated semi-implicitly using one-sided interpolation with the second-order accuracy. At each step, the unconditionally unique solvability of the unsymmetric linear system of equations in the proposed method is proved with a detailed discussion on the condition number. Numerically, the second-order accuracy in both time and space is verified. Using the proposed method, the inertial effect of ferromagnetics is observed in micromagnetics simulations at small timescales, in consistency with the hyperbolic property of the model at sub-picoseconds. For long time simulations, the results of the iLLG model are in nice agreements with those of the LLG model, in consistency with the parabolic feature of the iLLG model at larger timescales.

In easy-plane ferromagnets, we show that the interplay between a domain wall and a spin wave packet can be formulated as the collision of two massive particles with a gravity-like attraction. In the presence of magnetic dissipation, the domain wall mimics a particle subject to viscous friction, while the spin wave packet resembles a particle of variable mass. Due to attractive nature of the interaction, the domain wall acquires a backward displacement as a spin wave packet penetrating the domain wall, even though there is no change in momentum of the wave packet before and after penetration., Comment: 6 pages, 3 figures

Ferroelectric materials are spontaneous symmetry breaking systems characterized by ordered electric polarizations. Similar to its ferromagnetic counterpart, a ferroelectric domain wall can be regarded as a soft interface separating two different ferroelectric domains. Here we show that two bound state excitations of electric polarization (polar wave), or the vibration and breathing modes, can be hosted and propagate within the ferroelectric domain wall. Specially, the vibration polar wave has zero frequency gap, thus is constricted deeply inside ferroelectric domain wall, and can propagate even in the presence of local pinnings. The ferroelectric domain wall waveguide as demonstrated here, offers new paradigm in developing ferroelectric information processing units., Comment: 5 pages, 3 figures

Polarization, denoting the precession direction with respect to the background magnetization, is an intrinsic degree of freedom of spin wave. Using magnetic textures to control the spin wave polarization is fundamental and indispensable toward reprogrammable polarization-based magnonics. Here, we show that due to the intrinsic cubic anisotropy, a $90^\circ$ antiferromagnetic domain wall naturally acts as a spin wave retarder (wave-plate). Moreover, for a $90^\circ$ domain wall pair developed by introducing a second domain in a homogenous antiferromagnetic wire, the sign of retarding effect can be flipped by simply switching the direction of the intermediate domain., Comment: 7 pages, 5 figures

Lane change is a very demanding driving task and number of traffic accidents are induced by mistaken maneuvers. An automated lane change system has the potential to reduce driver workload and to improve driving safety. One challenge is how to improve driver acceptance on the automated system. From the viewpoint of human factors, an automated system with different styles would improve user acceptance as the drivers can adapt the style to different driving situations. This paper proposes a method to design different lane change styles in automated driving by analysis and modeling of truck driver behavior. A truck driving simulator experiment with 12 participants was conducted to identify the driver model parameters and three lane change styles were classified as the aggressive, medium, and conservative ones. The proposed automated lane change system was evaluated by another truck driving simulator experiment with the same 12 participants. Moreover, the effect of different driving styles on driver experience and acceptance was evaluated. The evaluation results demonstrate that the different lane change styles could be distinguished by the drivers; meanwhile, the three styles were overall evaluated as acceptable on safety issues and reliable by the human drivers. This study provides insight into designing the automated driving system with different driving styles and the findings can be applied to commercial automated trucks.

Spin wave, the collective excitation of magnetic order, is one of the fundamental angular momentum carriers in magnetic systems. Understanding the spin wave propagation in magnetic textures lies in the heart of developing pure magnetic information processing schemes. Here we show that the spin wave propagation across a chiral domain wall follows simple geometric trajectories, similar to the geometric optics. And the geometric behaviors are qualitatively different in normally magnetized film and tangentially magnetized film. We identify the lateral shift, refraction, and total reflection of spin wave across a ferromagnetic domain wall. Moreover, these geometric scattering phenomena become polarization-dependent in antiferromagnets, indicating the emergence of spin wave birefringence inside antiferromagnetic domain wall., Comment: 9 pages, 5 figures

Spin wave and magnetic texture are two elementary excitations in magnetic systems, and their interaction leads to rich magnetic phenomena. By describing the spin wave and the magnetic texture using their own collective coordinates, we find that they interact as classical particles traveling in mutual electromagnetic fields. Based on this unified collective coordinate model, we find that both skew scattering and side jump may occur as spin wave passing through magnetic textures. The skew scattering is associated with the magnetic topology of the texture, while the side jump is correlated to the total magnetization of the texture. We illustrate the concepts of skew scattering and side jump by investigating the spin wave trajectories across the topological magnetic Skyrmion and the topologically trivial magnetic bubble respectively., Comment: 5 pages, 3 figures

Quantum computing hardware has received world-wide attention and made considerable progress recently. YIG thin film have spin wave (magnon) modes with low dissipation and reliable control for quantum information processing. However, the coherent coupling between a quantum device and YIG thin film has yet been demonstrated. Here, we propose a scheme to achieve strong coupling between superconducting flux qubits and magnon modes in YIG thin film. Unlike the direct $\sqrt{N}$ enhancement factor in coupling to the Kittel mode or other spin ensembles, with N the total number of spins, an additional spatial dependent phase factor needs to be considered when the qubits are magnetically coupled with the magnon modes of finite wavelength. To avoid undesirable cancelation of coupling caused by the symmetrical boundary condition, a CoFeB thin layer is added to one side of the YIG thin film to break the symmetry. Our numerical simulation demonstrates avoided crossing and coherent transfer of quantum information between the flux qubits and the standing spin waves in YIG thin films. We show that the YIG thin film can be used as a tunable switch between two flux qubits, which have modified shape with small direct inductive coupling between them. Our results manifest that it is possible to couple flux qubits while suppressing undesirable cross-talk., Comment: 13 pages, 4 figures

Spin wave, the precession of magnetic order in magnetic materials, is a collective excitation that carries spin angular momentum. Similar to the acoustic or optical waves, the spin wave also possesses the polarization degree of freedom. Although such polarization degrees of freedom are frozen in ferromagnets, they are fully unlocked in antiferromagnets or ferrimagnets. Here we introduce the concept of magnetic gating and demonstrate a spin wave analog of the Datta-Das spin transistor in antiferromagnet. Utilizing the interplay between polarized spin wave and the antiferromagnetic domain walls, we propose a universal logic gate of pure magnetic nature, which realizes all Boolean operations in one single magnetic structure. We further construct a full functional 4-bit Arithmetic Logic Unit using only sixteen spin wave universal logic gates, operating in a weaving fashion as a Jacquard loom machine. The spin wave-based architecture proposed here also sets a model for the future energy efficient non-volatile computing, the distributed processing-in-memory computing, and the evolvable neuromorphic computing., Comment: 6 pages, 5 figures, 1 supplementary file. Supplemented movies can be requested from the authors