Searcharxiv⌕ Search

arXiv subjects

Hao Zhuang

Publications and source records attributed to Hao Zhuang.

At least 19 recordsLinked to original sources

RETR: A Structure-Preserving RGB-Event Transformer for Robust 3D Lane Detection

Robust 3D lane detection requires accurate metric road geometry recovery from visual inputs, yet conventional RGB-based approaches suffer from fragile lane evidence under adverse illumination, motion blur, and long-range perspective compression. Event cameras offer complementary high-temporal-resolution, high dynamic-range structural cues that are robust to these challenges, but their sparse, motion-dependent responses cannot be directly fused into consistent 3D geometry. In this paper, we present the first attempt to introduce event cameras into 3D lane detection. To enable systematic research, we first build two multimodal benchmarks with metric 3D lane annotations: DSEC-3DLD (real-world sequences) and Ev-OpenLane (large-scale simulated sequences). We further propose RETR, a structure preserving transformer that converts complementary RGB and event observations into coherent 3D lane geometry. RETR first aligns modality consistent evidence through Reciprocal Context Flow Fusion, then preserves thin and uncertain lane structures with Uncertainty Aware Structural Consolidation, and finally decodes ordered lane hypotheses using a Geometry State Decoder with proposal conditioned initialization, reference conditioned query evolution, and diversified geometric inquiry. Extensive experiments show that RETR achieves state-of-the-art performance on both benchmarks, especially showing strong improvements in challenging lighting and complex road geometry scenarios. Code and dataset will be available at https://github.com/duskswell/RETR.

cs.CV↗

Odd mapping cone Chern character

Our previous work established the mapping cone Chern character on the even-degree side, using the powers of the mapping cone curvature associated with the given mapping cone connection. In this paper, we carry out the odd-degree version using a trivial mapping cone connection and an automorphism on a trivial complex vector bundle. This version is based on the related odd-degree transgression.

math.DG↗

Analytic and topological realizations of the invariant Thom-Smale complex

For a Morse function, its associated Thom-Smale cochain complex admits an analytic realization initiated by Witten. However, due to the unboundedness of the eigenvalues of the deformed Hodge Laplacian along the critical submanifold, the analytic realization of the Thom-Smale complex associated with a Morse-Bott function is a long time open question. In this paper, we give an analytic realization in a case where a compact connected Lie group $G$ acts on a closed oriented manifold $M$. Our construction is not repeating the $G$-equivariant complex, but is actually a $G$-invariant complex computing the de Rham cohomology of $M$. The $G$-invariance is the key to resolve the unboundedness of the eigenvalues along the critical orbits. First, we fix a $G$-invariant Morse-Bott function $f$ on $M$ whose critical set consists of finitely many $G$-orbits, and whose Hessian in the normal direction of each critical orbit is nondegenerate. Second, on the topological side, we simplify the topological Thom-Smale cochain complex associated with $f$ into a $G$-invariant version given by $G$-invariant forms on critical orbits. Third, on the analytic side, we construct the $G$-invariant Witten instanton cochain complex after restricting the deformed Hodge Laplacian on $G$-invariant forms on $M$. As the main results, we prove that both $G$-invariant cochain complexes compute the Betti numbers of $M$, and that there is a cochain isomorphism between these two complexes. Thus, the $G$-invariant Witten instanton cochain complex is the analytic realization that we need.

math.DG↗

Instanton construction of the mapping cone Thom-Smale complex

The cup product structure on the topological side of the classical Thom-Smale complex leads to the topological side of the mapping cone Thom-Smale complex. Following the spirit of Witten's analytic Morse theory, we ask whether the mapping cone Thom-Smale complex has the analytic side. However, due to the missing cup product structure on the analytic side of the classical Thom-Smale complex, it seems that we can only have a hybrid analytic-topological construction of the mapping cone Thom-Smale complex. In this paper, we overcome the cup product issue and give the purely analytic construction of the mapping cone Thom-Smale complex. More precisely, for a Morse function with the transversality condition on a closed oriented Riemannian manifold, we construct an instanton cochain complex using the eigenspaces of the mapping cone Laplacian deformed by the Morse function and two parameters. One parameter gives the classical Witten deformation. The other parameter overcomes the cup product obstacle by suppressing the norm of the given differential form. As the main result, we prove that our instanton complex is cochain isomorphic to the topologically constructed mapping cone Thom-Smale complex, and therefore it is the purely analytic construction that we need.

math.DG↗

Emo-DVS: A Multimodal Benchmark for Privacy-Aware Emotion Recognition with Event Cameras

Emotion analysis is a fundamental task in computer vision, but its practical deployment remains constrained by the privacy risks inherent to conventional RGB cameras. Bio-inspired event cameras present a promising hardware-level solution because they capture asynchronous brightness changes, thereby reducing exposure of facial identity details while leveraging high dynamic range for robust perception under challenging illumination conditions. Despite these advantages, existing event-based methods struggle in complex real-world settings due to limited dataset scales, simple acquisition conditions, and reliance on single-modality visual cues. To address these, we establish a challenging tri-modal benchmark with event, audio, and text modalities and propose the Information-Guided Gated Fusion (IGF) framework, which first pre-trains an event encoder on the FAU subset of Emo-DVS to capture fine-grained facial dynamics, then employs adaptive modality gating to suppress modality-specific noise, and finally leverages mutual information maximization to align robust cross-modal representations. To alleviate data scarcity, we introduce Emo-DVS, the first large-scale event-based emotion analysis dataset, which couples dynamic illumination with the Facial Action Unit (FAU) subset and emotion subset. Extensive experiments demonstrate that IGF achieves state-of-the-art performance.

cs.CV↗

Bott connection in the mapping cone case

We generalize the concept of Bott connection to the de Rham mapping cone situation. As the main results, when the given closed smooth manifold is foliated, we have the mapping cone version of the Bott vanishing theorem and an adiabatic limit of a family of mapping cone connections on the orthogonal complement of the foliation. The second result shows the necessity of a refined metric if we want to see more interactions between geometry and topology in the mapping cone situation.

math.DG↗

Symplectic semi-characteristics

We study the symplectic semi-characteristic of a closed 4n-dimensional symplectic manifold. First, using the even-degree part of the primitive cohomology, we define the symplectic semi-characteristic. Second, using a vector field with nondegenerate zero points, we prove a counting formula for the symplectic semicharacteristic. As corollaries of the counting formula, we obtain a vanishing property and the fact that the definition of the symplectic semi-characteristic is independent of the choices of symplectic forms.

math.SG↗

A Serre type vanishing property of the twisted primitive cohomology

We prove a Serre type vanishing property for the twisted primitive cohomology of a symplectic manifold. It is based on Tseng and Zhou's vanishing property under the symplectic flatness. These vanishing properties emphasizes the necessity of the symplectic flatness when generalizing certain results from the sheaf cohomology in complex geometry to the primitive cohomology in symplectic geometry.

math.SG↗

Mapping cone Thom forms

For the de Rham mapping cone cochain complex induced by a smooth closed 2-form, we explicitly write down the associated mapping cone Thom form in the sense of Mathai-Quillen. Our construction uses the mapping cone covariant derivative, carrying the extra information brought by the 2-form. Our main tool is the Berezin integral. As the main result, we show that this Thom form is closed with respect to the mapping cone differentiation, its integration along the fiber is 1, and it satisfies the transgression formula.

math.DG↗

Transgression in the primitive cohomology

We study the Chern-Weil theory for the primitive cohomology of a symplectic manifold. First, given a symplectic manifold, we review the superbundle-valued forms on this manifold and prove a primitive version of the Bianchi identity. Second, as the main result, we prove a transgression formula associated with the boundary map of the primitive cohomology. Third, as an application of the main result, we introduce the concept of primitive characteristic classes and point out a further direction.

math.DG↗

A vanishing property about the 1-filtered cohomology groups of (4n+2)-dimensional closed symplectic manifolds

This note is a follow-up to our previous work arXiv:2505.14496. For any (4n+2)-dimensional closed symplectic manifold, we find that the dimension of the even-degree part of its 1-filtered cohomology is even, similar to the vanishing property of the classical Euler characteristic of an odd-dimensional closed manifold. We prove our result by constructing and then deforming a skew-adjoint operator. This process follows the methods in arXiv:2505.14496 but needs adjustments on signs and the power of the symplectic form.

math.SG↗

A High Compression Ratio Channel Multiplexing Method for Micro-pattern Gaseous Detectors

The demand for a large number of readout channels has been a limiting factor for the application of Micro-pattern Gaseous Detectors (MPGDs) in achieving higher spatial resolution and larger detection areas. This challenge is further compounded by issues related to system integration, power consumption, and cost efficiency. To address these challenges, this study proposes two novel multiplexing methods based on Eulerian circuits. Mathematical calculations indicate that with $n$ electronics channels, up to $n \times (n-1)/2 - (n - 2)/2 + 1$ detector channels can be read out, where $n$ is even. Three types of multiplexing circuits were designed, implemented, and tested in combination with Micromegas detectors. Experimental results demonstrate that, for a multiplexing circuit with a factor of 8, the spatial resolution remains comparable to the direct readout method, while achieving a detection efficiency exceeding 94\%. For a circuit with a multiplexing factor of 16, although the spatial resolution shows a slight degradation, the detection efficiency remains above 93.6\%. These results demonstrate that the proposed multiplexing methods can significantly reduce the number of readout channels while maintaining an acceptable level of spatial resolution and detection efficiency. These findings highlight the potential of the proposed multiplexing techniques for applications in fields requiring high-resolution and cost-effective detector systems, such as cosmic-ray muon imaging.

physics.ins-det↗

Nonlinear Motion-Guided and Spatio-Temporal Aware Network for Unsupervised Event-Based Optical Flow

Event cameras have the potential to capture continuous motion information over time and space, making them well-suited for optical flow estimation. However, most existing learning-based methods for event-based optical flow adopt frame-based techniques, ignoring the spatio-temporal characteristics of events. Additionally, these methods assume linear motion between consecutive events within the loss time window, which increases optical flow errors in long-time sequences. In this work, we observe that rich spatio-temporal information and accurate nonlinear motion between events are crucial for event-based optical flow estimation. Therefore, we propose E-NMSTFlow, a novel unsupervised event-based optical flow network focusing on long-time sequences. We propose a Spatio-Temporal Motion Feature Aware (STMFA) module and an Adaptive Motion Feature Enhancement (AMFE) module, both of which utilize rich spatio-temporal information to learn spatio-temporal data associations. Meanwhile, we propose a nonlinear motion compensation loss that utilizes the accurate nonlinear motion between events to improve the unsupervised learning of our network. Extensive experiments demonstrate the effectiveness and superiority of our method. Remarkably, our method ranks first among unsupervised learning methods on the MVSEC and DSEC-Flow datasets. Our project page is available at https://wynelio.github.io/E-NMSTFlow.

cs.CV↗

Kervaire semi-characteristics in KK-theory and an Atiyah type vanishing theorem

On (4n + 1)-dimensional (noncompact) manifolds admitting proper cocompact Lie group actions, we explore the analytic and topological sides of Kervaire semi-characteristics. The analytic side puts together two interpretations, one via assembly maps, and the other via dimensions of kernels. The topological side is ensured by the proper cocompact version of the Hodge theorem. The two sides coincide and admit an Atiyah type vanishing theorem.

math.DG↗

EV-MGDispNet: Motion-Guided Event-Based Stereo Disparity Estimation Network with Left-Right Consistency

Event cameras have the potential to revolutionize the field of robot vision, particularly in areas like stereo disparity estimation, owing to their high temporal resolution and high dynamic range. Many studies use deep learning for event camera stereo disparity estimation. However, these methods fail to fully exploit the temporal information in the event stream to acquire clear event representations. Additionally, there is room for further reduction in pixel shifts in the feature maps before constructing the cost volume. In this paper, we propose EV-MGDispNet, a novel event-based stereo disparity estimation method. Firstly, we propose an edge-aware aggregation (EAA) module, which fuses event frames and motion confidence maps to generate a novel clear event representation. Then, we propose a motion-guided attention (MGA) module, where motion confidence maps utilize deformable transformer encoders to enhance the feature map with more accurate edges. Finally, we also add a census left-right consistency loss function to enhance the left-right consistency of stereo event representation. Through conducting experiments within challenging real-world driving scenarios, we validate that our method outperforms currently known state-of-the-art methods in terms of mean absolute error (MAE) and root mean square error (RMSE) metrics.

cs.CV↗

EV-MGRFlowNet: Motion-Guided Recurrent Network for Unsupervised Event-based Optical Flow with Hybrid Motion-Compensation Loss

Event cameras offer promising properties, such as high temporal resolution and high dynamic range. These benefits have been utilized into many machine vision tasks, especially optical flow estimation. Currently, most existing event-based works use deep learning to estimate optical flow. However, their networks have not fully exploited prior hidden states and motion flows. Additionally, their supervision strategy has not fully leveraged the geometric constraints of event data to unlock the potential of networks. In this paper, we propose EV-MGRFlowNet, an unsupervised event-based optical flow estimation pipeline with motion-guided recurrent networks using a hybrid motion-compensation loss. First, we propose a feature-enhanced recurrent encoder network (FERE-Net) which fully utilizes prior hidden states to obtain multi-level motion features. Then, we propose a flow-guided decoder network (FGD-Net) to integrate prior motion flows. Finally, we design a hybrid motion-compensation loss (HMC-Loss) to strengthen geometric constraints for the more accurate alignment of events. Experimental results show that our method outperforms the current state-of-the-art (SOTA) method on the MVSEC dataset, with an average reduction of approximately 22.71% in average endpoint error (AEE). To our knowledge, our method ranks first among unsupervised learning-based methods.

cs.CV↗

FE-Fusion-VPR: Attention-based Multi-Scale Network Architecture for Visual Place Recognition by Fusing Frames and Events

Traditional visual place recognition (VPR), usually using standard cameras, is easy to fail due to glare or high-speed motion. By contrast, event cameras have the advantages of low latency, high temporal resolution, and high dynamic range, which can deal with the above issues. Nevertheless, event cameras are prone to failure in weakly textured or motionless scenes, while standard cameras can still provide appearance information in this case. Thus, exploiting the complementarity of standard cameras and event cameras can effectively improve the performance of VPR algorithms. In the paper, we propose FE-Fusion-VPR, an attention-based multi-scale network architecture for VPR by fusing frames and events. First, the intensity frame and event volume are fed into the two-stream feature extraction network for shallow feature fusion. Next, the three-scale features are obtained through the multi-scale fusion network and aggregated into three sub-descriptors using the VLAD layer. Finally, the weight of each sub-descriptor is learned through the descriptor re-weighting network to obtain the final refined descriptor. Experimental results show that on the Brisbane-Event-VPR and DDD20 datasets, the Recall@1 of our FE-Fusion-VPR is 29.26% and 33.59% higher than Event-VPR and Ensemble-EventVPR, and is 7.00% and 14.15% higher than MultiRes-NetVLAD and NetVLAD. To our knowledge, this is the first end-to-end network that goes beyond the existing event-based and frame-based SOTA methods to fuse frame and events directly for VPR.

cs.CV↗

Neuro-Planner: A 3D Visual Navigation Method for MAV with Depth Camera based on Neuromorphic Reinforcement Learning

Traditional visual navigation methods of micro aerial vehicle (MAV) usually calculate a passable path that satisfies the constraints depending on a prior map. However, these methods have issues such as high demand for computing resources and poor robustness in face of unfamiliar environments. Aiming to solve the above problems, we propose a neuromorphic reinforcement learning method (Neuro-Planner) that combines spiking neural network (SNN) and deep reinforcement learning (DRL) to realize MAV 3D visual navigation with depth camera. Specifically, we design spiking actor network based on two-state LIF (TS-LIF) neurons and its encoding-decoding schemes for efficient inference. Then our improved hybrid deep deterministic policy gradient (HDDPG) and TS-LIF-based spatio-temporal back propagation (STBP) algorithms are used as the training framework for actor-critic network architecture. To verify the effectiveness of the proposed Neuro-Planner, we carry out detailed comparison experiments with various SNN training algorithm (STBP, BPTT and SLAYER) in the software-in-the-loop (SITL) simulation framework. The navigation success rate of our HDDPG-STBP is 4.3\% and 5.3\% higher than that of the original DDPG in the two evaluation environments. To the best of our knowledge, this is the first work combining neuromorphic computing and deep reinforcement learning for MAV 3D visual navigation task.

cs.RO↗