arXiv subjects

Xingang Wang

Publications and source records attributed to Xingang Wang.

At least 55 records · Page 3Linked to original sources

arXiv2023

DiffBEV: Conditional Diffusion Model for Bird's Eye View Perception

BEV perception is of great importance in the field of autonomous driving, serving as the cornerstone of planning, controlling, and motion prediction. The quality of the BEV feature highly affects the performance of BEV perception. However, taking the noises in camera parameters and LiDAR scans into consideration, we usually obtain BEV representation with harmful noises. Diffusion models naturally have the ability to denoise noisy samples to the ideal data, which motivates us to utilize the diffusion model to get a better BEV representation. In this work, we propose an end-to-end framework, named DiffBEV, to exploit the potential of diffusion model to generate a more comprehensive BEV representation. To the best of our knowledge, we are the first to apply diffusion model to BEV perception. In practice, we design three types of conditions to guide the training of the diffusion model which denoises the coarse samples and refines the semantic feature in a progressive way. What's more, a cross-attention module is leveraged to fuse the context of BEV feature and the semantic content of conditional diffusion model. DiffBEV achieves a 25.9% mIoU on the nuScenes dataset, which is 6.2% higher than the best-performing existing approach. Quantitative and qualitative results on multiple benchmarks demonstrate the effectiveness of DiffBEV in BEV semantic segmentation and 3D object detection tasks. The code will be available soon.

cs.CV↗

arXiv2023

OpenOccupancy: A Large Scale Benchmark for Surrounding Semantic Occupancy Perception

Semantic occupancy perception is essential for autonomous driving, as automated vehicles require a fine-grained perception of the 3D urban structures. However, existing relevant benchmarks lack diversity in urban scenes, and they only evaluate front-view predictions. Towards a comprehensive benchmarking of surrounding perception algorithms, we propose OpenOccupancy, which is the first surrounding semantic occupancy perception benchmark. In the OpenOccupancy benchmark, we extend the large-scale nuScenes dataset with dense semantic occupancy annotations. Previous annotations rely on LiDAR points superimposition, where some occupancy labels are missed due to sparse LiDAR channels. To mitigate the problem, we introduce the Augmenting And Purifying (AAP) pipeline to ~2x densify the annotations, where ~4000 human hours are involved in the labeling process. Besides, camera-based, LiDAR-based and multi-modal baselines are established for the OpenOccupancy benchmark. Furthermore, considering the complexity of surrounding occupancy perception lies in the computational burden of high-resolution 3D predictions, we propose the Cascade Occupancy Network (CONet) to refine the coarse prediction, which relatively enhances the performance by ~30% than the baseline. We hope the OpenOccupancy benchmark will boost the development of surrounding occupancy perception algorithms.

cs.CV↗

arXiv2023

Inferring synchronizability of networked heterogenous oscillators with machine learning

In the study of network synchronization, an outstanding question of both theoretical and practical significance is how to allocate a given set of heterogenous oscillators on a complex network in order for improving the synchronization performance. Whereas methods have been proposed to address this question in literature, the methods are based on accurate models describing the system dynamics, which, however, are normally unavailable in realistic situations. Here we show that this question can be addressed by the model-free technique of feed-forward neural network (FNN) in machine learning. Specifically, we measure the synchronization performance of a number of allocation schemes and use the measured data to train a machine. It is found that the trained machine is able to not only infer the synchronization performance of any new allocation scheme, but also find from a huge amount of candidates the optimal allocation scheme for synchronization.

nlin.AO↗

arXiv2023

Breathing cluster in complex neuron-astrocyte networks

Brain activities are featured by spatially distributed neural clusters of coherent firings and a spontaneous switching of the clusters between the synchrony and asynchrony states. Evidences from {\it in vivo} experiments suggest that astrocytes, a type of glial cell regarded previously as providing only structural and metabolic supports to neurons, participate actively in brain functions and play a crucial role in regulating the neural firing activities, yet the mechanism remains unknown. Introducing astrocyte as a reservoir of the glutamate released from neuron synapses, here we propose the model of complex neuron-astrocyte network and employ it to explore the roles of astrocyte in regulating the synchronization behaviors of networked neurons. It is found that a fraction of neurons on the network can be synchronized as a cluster, while the remaining neurons are kept as desynchronized. Moreover, during the course of network evolution, the cluster is switching between the synchrony and asynchrony states intermittently, henceforth the phenomenon of ``breathing cluster". By the method of symmetry-based analysis, we conduct a theoretical investigation on the stability of the cluster and the mechanism generating the breathing activities. It is revealed that the contents of the cluster are determined by the network symmetry and the breathing activities are due to the interplay between the neural network and the astrocyte. The breathing phenomenon is demonstrated in network models of different structures and neural dynamics. The studies give insights into the cellular mechanism of astrocytes in regulating neural activities, and shed lights onto the spontaneous state switching of the neocortex.

q-bio.NC↗

arXiv2022

Are We Ready for Vision-Centric Driving Streaming Perception? The ASAP Benchmark

In recent years, vision-centric perception has flourished in various autonomous driving tasks, including 3D detection, semantic map construction, motion forecasting, and depth estimation. Nevertheless, the latency of vision-centric approaches is too high for practical deployment (e.g., most camera-based 3D detectors have a runtime greater than 300ms). To bridge the gap between ideal research and real-world applications, it is necessary to quantify the trade-off between performance and efficiency. Traditionally, autonomous-driving perception benchmarks perform the offline evaluation, neglecting the inference time delay. To mitigate the problem, we propose the Autonomous-driving StreAming Perception (ASAP) benchmark, which is the first benchmark to evaluate the online performance of vision-centric perception in autonomous driving. On the basis of the 2Hz annotated nuScenes dataset, we first propose an annotation-extending pipeline to generate high-frame-rate labels for the 12Hz raw images. Referring to the practical deployment, the Streaming Perception Under constRained-computation (SPUR) evaluation protocol is further constructed, where the 12Hz inputs are utilized for streaming evaluation under the constraints of different computational resources. In the ASAP benchmark, comprehensive experiment results reveal that the model rank alters under different constraints, suggesting that the model latency and computation budget should be considered as design choices to optimize the practical deployment. To facilitate further research, we establish baselines for camera-based streaming 3D detection, which consistently enhance the streaming performance across various hardware. ASAP project page: https://github.com/JeffWang987/ASAP.

cs.CV↗

arXiv2022

Multi-Granularity Distillation Scheme Towards Lightweight Semi-Supervised Semantic Segmentation

Albeit with varying degrees of progress in the field of Semi-Supervised Semantic Segmentation, most of its recent successes are involved in unwieldy models and the lightweight solution is still not yet explored. We find that existing knowledge distillation techniques pay more attention to pixel-level concepts from labeled data, which fails to take more informative cues within unlabeled data into account. Consequently, we offer the first attempt to provide lightweight SSSS models via a novel multi-granularity distillation (MGD) scheme, where multi-granularity is captured from three aspects: i) complementary teacher structure; ii) labeled-unlabeled data cooperative distillation; iii) hierarchical and multi-levels loss setting. Specifically, MGD is formulated as a labeled-unlabeled data cooperative distillation scheme, which helps to take full advantage of diverse data characteristics that are essential in the semi-supervised setting. Image-level semantic-sensitive loss, region-level content-aware loss, and pixel-level consistency loss are set up to enrich hierarchical distillation abstraction via structurally complementary teachers. Experimental results on PASCAL VOC2012 and Cityscapes reveal that MGD can outperform the competitive approaches by a large margin under diverse partition protocols. For example, the performance of ResNet-18 and MobileNet-v2 backbone is boosted by 11.5% and 4.6% respectively under 1/16 partition protocol on Cityscapes. Although the FLOPs of the model backbone is compressed by 3.4-5.3x (ResNet-18) and 38.7-59.6x (MobileNetv2), the model manages to achieve satisfactory segmentation results.

cs.CV↗

arXiv2022

Crafting Monocular Cues and Velocity Guidance for Self-Supervised Multi-Frame Depth Learning

Self-supervised monocular methods can efficiently learn depth information of weakly textured surfaces or reflective objects. However, the depth accuracy is limited due to the inherent ambiguity in monocular geometric modeling. In contrast, multi-frame depth estimation methods improve the depth accuracy thanks to the success of Multi-View Stereo (MVS), which directly makes use of geometric constraints. Unfortunately, MVS often suffers from texture-less regions, non-Lambertian surfaces, and moving objects, especially in real-world video sequences without known camera motion and depth supervision. Therefore, we propose MOVEDepth, which exploits the MOnocular cues and VElocity guidance to improve multi-frame Depth learning. Unlike existing methods that enforce consistency between MVS depth and monocular depth, MOVEDepth boosts multi-frame depth learning by directly addressing the inherent problems of MVS. The key of our approach is to utilize monocular depth as a geometric priority to construct MVS cost volume, and adjust depth candidates of cost volume under the guidance of predicted camera velocity. We further fuse monocular depth and MVS depth by learning uncertainty in the cost volume, which results in a robust depth estimation against ambiguity in multi-view geometry. Extensive experiments show MOVEDepth achieves state-of-the-art performance: Compared with Monodepth2 and PackNet, our method relatively improves the depth accuracy by 20\% and 19.8\% on the KITTI benchmark. MOVEDepth also generalizes to the more challenging DDAD benchmark, relatively outperforming ManyDepth by 7.2\%. The code is available at https://github.com/JeffWang987/MOVEDepth.

cs.CV↗

arXiv2022

MVSTER: Epipolar Transformer for Efficient Multi-View Stereo

Learning-based Multi-View Stereo (MVS) methods warp source images into the reference camera frustum to form 3D volumes, which are fused as a cost volume to be regularized by subsequent networks. The fusing step plays a vital role in bridging 2D semantics and 3D spatial associations. However, previous methods utilize extra networks to learn 2D information as fusing cues, underusing 3D spatial correlations and bringing additional computation costs. Therefore, we present MVSTER, which leverages the proposed epipolar Transformer to learn both 2D semantics and 3D spatial associations efficiently. Specifically, the epipolar Transformer utilizes a detachable monocular depth estimator to enhance 2D semantics and uses cross-attention to construct data-dependent 3D associations along epipolar line. Additionally, MVSTER is built in a cascade structure, where entropy-regularized optimal transport is leveraged to propagate finer depth estimations in each stage. Extensive experiments show MVSTER achieves state-of-the-art reconstruction performance with significantly higher efficiency: Compared with MVSNet and CasMVSNet, our MVSTER achieves 34% and 14% relative improvements on the DTU benchmark, with 80% and 51% relative reductions in running time. MVSTER also ranks first on Tanks&Temples-Advanced among all published works. Code is released at https://github.com/JeffWang987.

cs.CV↗

arXiv2022

HFT: Lifting Perspective Representations via Hybrid Feature Transformation

Autonomous driving requires accurate and detailed Bird's Eye View (BEV) semantic segmentation for decision making, which is one of the most challenging tasks for high-level scene perception. Feature transformation from frontal view to BEV is the pivotal technology for BEV semantic segmentation. Existing works can be roughly classified into two categories, i.e., Camera model-Based Feature Transformation (CBFT) and Camera model-Free Feature Transformation (CFFT). In this paper, we empirically analyze the vital differences between CBFT and CFFT. The former transforms features based on the flat-world assumption, which may cause distortion of regions lying above the ground plane. The latter is limited in the segmentation performance due to the absence of geometric priors and time-consuming computation. In order to reap the benefits and avoid the drawbacks of CBFT and CFFT, we propose a novel framework with a Hybrid Feature Transformation module (HFT). Specifically, we decouple the feature maps produced by HFT for estimating the layout of outdoor scenes in BEV. Furthermore, we design a mutual learning scheme to augment hybrid transformation by applying feature mimicking. Notably, extensive experiments demonstrate that with negligible extra overhead, HFT achieves a relative improvement of 13.3% on the Argoverse dataset and 16.8% on the KITTI 3D Object datasets compared to the best-performing existing method. The codes are available at https://github.com/JiayuZou2020/HFT.

cs.CV↗

arXiv2021

Activation Modulation and Recalibration Scheme for Weakly Supervised Semantic Segmentation

Image-level weakly supervised semantic segmentation (WSSS) is a fundamental yet challenging computer vision task facilitating scene understanding and automatic driving. Most existing methods resort to classification-based Class Activation Maps (CAMs) to play as the initial pseudo labels, which tend to focus on the discriminative image regions and lack customized characteristics for the segmentation task. To alleviate this issue, we propose a novel activation modulation and recalibration (AMR) scheme, which leverages a spotlight branch and a compensation branch to obtain weighted CAMs that can provide recalibration supervision and task-specific concepts. Specifically, an attention modulation module (AMM) is employed to rearrange the distribution of feature importance from the channel-spatial sequential perspective, which helps to explicitly model channel-wise interdependencies and spatial encodings to adaptively modulate segmentation-oriented activation responses. Furthermore, we introduce a cross pseudo supervision for dual branches, which can be regarded as a semantic similar regularization to mutually refine two branches. Extensive experiments show that AMR establishes a new state-of-the-art performance on the PASCAL VOC 2012 dataset, surpassing not only current methods trained with the image-level of supervision but also some methods relying on stronger supervision, such as saliency label. Experiments also reveal that our scheme is plug-and-play and can be incorporated with other approaches to boost their performance.

cs.CV↗

arXiv2021

Learning the dynamics of coupled oscillators from transients

Whereas the importance of transient dynamics to the functionality and management of complex systems has been increasingly recognized, most of the studies are based on models. Yet in realistic situations the models are often unknown and what available are only measured time series. Meanwhile, many real-world systems are dynamically stable, in the sense that the systems return to their equilibria in a short time after perturbations. This increases further the difficulty of dynamics analysis, as many information of the system dynamics are lost once the system is settled onto the equilibrium states. The question we ask is: given the transient time series of a complex dynamical system measured in the stable regime, can we infer from the data some properties of the system dynamics and make predictions, e.g., predicting the critical point where the equilibrium state becomes unstable? We show that for the typical transitions in system of coupled oscillators, including quorum sensing, amplitude death and complete synchronization, this question can be addressed by the technique of reservoir computing in machine learning. More specifically, by the transient series acquired at several states in the stable regime, we demonstrate that the trained machine is able to predict accurately not only the transient behaviors of the system in the stable regime, but also the critical point where the stable state becomes unstable. Considering the ubiquitous existence of transient activities in natural and man-made systems, the findings may have broad applications.

nlin.AO↗

arXiv2021

Criticality in Reservoir Computer of Coupled Phase Oscillators

Accumulating evidences show that the cerebral cortex is operating near a critical state featured by power-law size distribution of neural avalanche activities, yet evidence of this critical state in artificial neural networks mimicking the cerebral cortex is lacking. Here we design an artificial neural network of coupled phase oscillators and, by the technique of reservoir computing in machine learning, train it for predicting chaos. It is found that when the machine is properly trained, oscillators in the reservoir are synchronized into clusters whose sizes follow a power-law distribution. This feature, however, is absent when the machine is poorly trained. Additionally, it is found that despite the synchronization degree of the original network, once properly trained, the reservoir network is always developed to the same critical state, exemplifying the "attractor" nature of this state in machine learning. The generality of the results is verified in different reservoir models and by different target systems, and it is found that the scaling exponent of the distribution is independent on the reservoir details and the bifurcation parameter of the target system, but is modified when the dynamics of the target system is changed to a different type. The findings shed lights on the nature of machine learning, and are helpful to the design of high-performance machine in physical systems.

q-bio.NC↗

arXiv2021

Learning Hamiltonian dynamics by reservoir computer

Reconstructing the KAM dynamics diagram of Hamiltonian system from the time series of a limited number of parameters is an outstanding question in nonlinear science, especially when the Hamiltonian governing the system dynamics are unknown. Here, we demonstrate that this question can be addressed by the machine learning approach knowing as reservoir computer (RC). Specifically, we show that without prior knowledge about the Hamilton's equations of motion, the trained RC is able to not only predict the short-term evolution of the system state, but also replicate the long-term ergodic properties of the system dynamics. Furthermore, by the architecture of parameter-aware RC, we also show that the RC trained by the time series acquired at a handful parameters is able to reconstruct the entire KAM dynamics diagram with a high precision by tuning a control parameter externally. The feasibility and efficiency of the learning techniques are demonstrated in two classical nonlinear Hamiltonian systems, namely the double-pendulum oscillator and the standard map. Our study indicates that, as a complex dynamical system, RC is able to learn from data the Hamiltonian.

eess.SP↗

arXiv2021

Anticipating synchronization with machine learning

In applications of dynamical systems, situations can arise where it is desired to predict the onset of synchronization as it can lead to characteristic and significant changes in the system performance and behaviors, for better or worse. In experimental and real settings, the system equations are often unknown, raising the need to develop a prediction framework that is model free and fully data driven. We contemplate that this challenging problem can be addressed with machine learning. In particular, exploiting reservoir computing or echo state networks, we devise a "parameter-aware" scheme to train the neural machine using asynchronous time series, i.e., in the parameter regime prior to the onset of synchronization. A properly trained machine will possess the power to predict the synchronization transition in that, with a given amount of parameter drift, whether the system would remain asynchronous or exhibit synchronous dynamics can be accurately anticipated. We demonstrate the machine-learning based framework using representative chaotic models and small network systems that exhibit continuous (second-order) or abrupt (first-order) transitions. A remarkable feature is that, for a network system exhibiting an explosive (first-order) transition and a hysteresis loop in synchronization, the machine learning scheme is capable of accurately predicting these features, including the precise locations of the transition points associated with the forward and backward transition paths.

nlin.AO↗

arXiv2020

ResizeMix: Mixing Data with Preserved Object Information and True Labels

Data augmentation is a powerful technique to increase the diversity of data, which can effectively improve the generalization ability of neural networks in image recognition tasks. Recent data mixing based augmentation strategies have achieved great success. Especially, CutMix uses a simple but effective method to improve the classifiers by randomly cropping a patch from one image and pasting it on another image. To further promote the performance of CutMix, a series of works explore to use the saliency information of the image to guide the mixing. We systematically study the importance of the saliency information for mixing data, and find that the saliency information is not so necessary for promoting the augmentation performance. Furthermore, we find that the cutting based data mixing methods carry two problems of label misallocation and object information missing, which cannot be resolved simultaneously. We propose a more effective but very easily implemented method, namely ResizeMix. We mix the data by directly resizing the source image to a small patch and paste it on another image. The obtained patch preserves more substantial object information compared with conventional cut-based methods. ResizeMix shows evident advantages over CutMix and the saliency-guided methods on both image classification and object detection tasks without additional computation cost, which even outperforms most costly search-based automatic augmentation methods.

cs.CV↗

arXiv2020

Spiral wave chimeras in reaction-diffusion systems: phenomenon, mechanism and transitions

Spiral wave chimeras (SWCs), which combine the features of spiral waves and chimera states, are a new type of dynamical patterns emerged in spatiotemporal systems due to the spontaneous symmetry breaking of the system dynamics. In generating SWC, the conventional wisdom is that the dynamical elements should be coupled in a nonlocal fashion. For this reason, it is commonly believed that SWC is excluded from the general reaction-diffusion (RD) systems possessing only local couplings. Here, by an experimentally feasible model of three-component FitzHugh-Nagumo-type RD system, we demonstrate that, even though the system elements are locally coupled, stable SWCs can still be observed in a wide region in the parameter space. The properties of SWCs are explored, and the underlying mechanisms are analyzed from the point view of coupled oscillators. Transitions from SWC to incoherent states are also investigated, and it is found that SWCs are typically destabilized in two scenarios, namely core breakup and core expansion. The former is characterized by a continuous breakup of the single asynchronous core into a number of small asynchronous cores, whereas the latter is featured by the continuous expansion of the single asynchronous core to the whole space. Remarkably, in the scenario of core expansion, the system may develop into an intriguing state in which regular spiral waves are embedded in a completely disordered background. This state, which is named shadowed spirals, manifests from a new perspective the coexistence of incoherent and coherent states in spatiotemporal systems, generalizing therefore the traditional concept of chimera states. Our studies provide an affirmative answer to the observation of SWCs in RD systems, and pave a way to the realization of SWCs in experiments.

nlin.PS↗

arXiv2020

Synchronization within synchronization: transients and intermittency in ecological networks

Transients are fundamental to ecological systems with significant implications to management, conservation, and biological control. We uncover a type of transient synchronization behavior in spatial ecological networks whose local dynamics are of the chaotic, predator-prey type. In the parameter regime where there is phase synchronization among all the patches, complete synchronization (i.e., synchronization in both phase and amplitude) can arise in certain pairs of patches as determined by the network symmetry - henceforth the phenomenon of "synchronization within synchronization." Distinct patterns of complete synchronization coexist but, due to intrinsic instability or noise, each pattern is a transient and there is random, intermittent switching among the patterns in the course of time evolution. The probability distribution of the transient time is found to follow an algebraic scaling law with a divergent average transient lifetime. Based on symmetry considerations, we develop a stability analysis to understand these phenomena. The general principle of symmetry can also be exploited to explain previously discovered, counterintuitive synchronization behaviors in ecological networks.

q-bio.PE↗

arXiv2020

Transfer learning of chaotic systems

Can a neural network trained by the time series of system A be used to predict the evolution of system B? This problem, knowing as transfer learning in a broad sense, is of great importance in machine learning and data mining, yet has not been addressed for chaotic systems. Here we investigate transfer learning of chaotic systems from the perspective of synchronization-based state inference, in which a reservoir computer trained by chaotic system A is used to infer the unmeasured variables of chaotic system B, while A is different from B in either parameter or dynamics. It is found that if systems A and B are different in parameter, the reservoir computer can be well synchronized to system B. However, if systems A and B are different in dynamics, the reservoir computer fails to synchronize with system B in general. Knowledge transfer along a chain of coupled reservoir computers is also studied, and it is found that, although the reservoir computers are trained by different systems, the unmeasured variables of the driving system can be successfully inferred by the remote reservoir computer. Finally, by an experiment of chaotic pendulum, we show that the knowledge learned from the modeling system can be used to predict the evolution of the experimental system.

cs.NE↗