arXiv subjects

Xiao Sun

Publications and source records attributed to Xiao Sun.

At least 91 records · Page 5Linked to original sources

arXiv2024

Within the Dynamic Context: Inertia-aware 3D Human Modeling with Pose Sequence

Neural rendering techniques have significantly advanced 3D human body modeling. However, previous approaches often overlook dynamics induced by factors such as motion inertia, leading to challenges in scenarios like abrupt stops after rotation, where the pose remains static while the appearance changes. This limitation arises from reliance on a single pose as conditional input, resulting in ambiguity in mapping one pose to multiple appearances. In this study, we elucidate that variations in human appearance depend not only on the current frame's pose condition but also on past pose states. Therefore, we introduce Dyco, a novel method utilizing the delta pose sequence representation for non-rigid deformations and canonical space to effectively model temporal appearance variations. To prevent a decrease in the model's generalization ability to novel poses, we further propose low-dimensional global context to reduce unnecessary inter-body part dependencies and a quantization operation to mitigate overfitting of the delta pose sequence by the model. To validate the effectiveness of our approach, we collected a novel dataset named I3D-Human, with a focus on capturing temporal changes in clothing appearance under approximate poses. Through extensive experiments on both I3D-Human and existing datasets, our approach demonstrates superior qualitative and quantitative performance. In addition, our inertia-aware 3D human method can unprecedentedly simulate appearance changes caused by inertia at different velocities.

cs.CV↗

arXiv2024

A Deep Learning-Based Target Radial Length Estimation Method through HRRP Sequence

This paper introduces an innovative deep learning-based method for end-to-end target radial length estimation from HRRP (High Resolution Range Profile) sequences. Firstly, the HRRP sequences are normalized and transformed into GAF (Gram Angular Field) images to effectively capture and utilize the temporal information. Subsequently, these GAF images serve as the input for a pretrained ResNet-101 model, which is then fine-tuned for target radial length estimation. The simulation results show that compared to traditional threshold method and simple networks e.g. one-dimensional CNN (Convolutional Neural Network), the proposed method demonstrates superior noise resistance and higher accuracy under low SNR (Signal-to-Noise Ratio) conditions.

eess.SP↗

arXiv2024

Narrow Linewidth Laser Based on Extended Topological Interface States in One-Dimensional Photonic Crystals

Recent advances in topological one-dimensional photonic crystal concepts have enabled the development of robust light-emitting devices by incorporating a topological interface state (TIS) at the cavity center. In this study, we theoretically and experimentally demonstrate a one-dimensional TIS-extended photonic crystal (1D-TISE-PC) structure. By integrating a linearly dispersive zero-index one-dimensional photonic crystal structure with a four-phase shift sampled grating, photons propagate along the cavity without phase differences, enhancing the robustness to material variations and extending the TIS. Our findings indicate that extending the TIS promotes a more uniform photon distribution along the laser cavity and mitigates the spatial hole burning (SHB) effect. We fabricated and characterized a 1550 nm sidewall 1D-TISE-PC semiconductor laser, achieving stable single-mode operation across a wide current range from 60 to 420 mA, with a side-mode suppression ratio of 50 dB. The 1D-TISE-PC structure exhibited a linewidth narrowing effect to approximately 150 kHz Lorentzian linewidth. Utilizing reconstruction equivalent-chirp technology for the 4PS sampled grating enabled precise wavelength control in 1D-TISE-PC laser arrays, achieving a wavelength spacing of 0.796 nm +- 0.003 nm. We show that the TIS still exists in the TISE cavity and topological protection is preserved. Its mode extension characteristics mitigate the SHB so narrows the linewidth. We argue that the design simplicity and improvement of the fabrication tolerance make this architecture suitable for high-power and narrow-linewidth semiconductor lasers development.

physics.optics↗

arXiv2024

Mask as Supervision: Leveraging Unified Mask Information for Unsupervised 3D Pose Estimation

Automatic estimation of 3D human pose from monocular RGB images is a challenging and unsolved problem in computer vision. In a supervised manner, approaches heavily rely on laborious annotations and present hampered generalization ability due to the limited diversity of 3D pose datasets. To address these challenges, we propose a unified framework that leverages mask as supervision for unsupervised 3D pose estimation. With general unsupervised segmentation algorithms, the proposed model employs skeleton and physique representations that exploit accurate pose information from coarse to fine. Compared with previous unsupervised approaches, we organize the human skeleton in a fully unsupervised way which enables the processing of annotation-free data and provides ready-to-use estimation results. Comprehensive experiments demonstrate our state-of-the-art pose estimation performance on Human3.6M and MPI-INF-3DHP datasets. Further experiments on in-the-wild datasets also illustrate the capability to access more data to boost our model. Code will be available at https://github.com/Charrrrrlie/Mask-as-Supervision.

cs.CV↗

arXiv2024

Joint Spatial-Temporal Modeling and Contrastive Learning for Self-supervised Heart Rate Measurement

This paper briefly introduces the solutions developed by our team, HFUT-VUT, for Track 1 of self-supervised heart rate measurement in the 3rd Vision-based Remote Physiological Signal Sensing (RePSS) Challenge hosted at IJCAI 2024. The goal is to develop a self-supervised learning algorithm for heart rate (HR) estimation using unlabeled facial videos. To tackle this task, we present two self-supervised HR estimation solutions that integrate spatial-temporal modeling and contrastive learning, respectively. Specifically, we first propose a non-end-to-end self-supervised HR measurement framework based on spatial-temporal modeling, which can effectively capture subtle rPPG clues and leverage the inherent bandwidth and periodicity characteristics of rPPG to constrain the model. Meanwhile, we employ an excellent end-to-end solution based on contrastive learning, aiming to generalize across different scenarios from complementary perspectives. Finally, we combine the strengths of the above solutions through an ensemble strategy to generate the final predictions, leading to a more accurate HR estimation. As a result, our solutions achieved a remarkable RMSE score of 8.85277 on the test dataset, securing \textbf{2nd place} in Track 1 of the challenge.

cs.CV↗

arXiv2024

Tightly-Coupled VLP/INS Integrated Navigation by Inclination Estimation and Blockage Handling

Visible Light Positioning (VLP) has emerged as a promising technology capable of delivering indoor localization with high accuracy. In VLP systems that use Photodiodes (PDs) as light receivers, the Received Signal Strength (RSS) is affected by the incidence angle of light, making the inclination of PDs a critical parameter in the positioning model. Currently, most studies assume the inclination to be constant, limiting the applications and positioning accuracy. Additionally, light blockages may severely interfere with the RSS measurements but the literature has not explored blockage detection in real-world experiments. To address these problems, we propose a tightly coupled VLP/INS (Inertial Navigation System) integrated navigation system that uses graph optimization to account for varying PD inclinations and VLP blockages. We also discussed the possibility of simultaneously estimating the robot's pose and the locations of some unknown LEDs. Simulations and two groups of real-world experiments demonstrate the efficiency of our approach, achieving an average positioning accuracy of 10 cm during movement and inclination accuracy within 1 degree despite inclination changes and blockages.

cs.RO↗

arXiv2024

DiffBody: Human Body Restoration by Imagining with Generative Diffusion Prior

Human body restoration plays a vital role in various applications related to the human body. Despite recent advances in general image restoration using generative models, their performance in human body restoration remains mediocre, often resulting in foreground and background blending, over-smoothing surface textures, missing accessories, and distorted limbs. Addressing these challenges, we propose a novel approach by constructing a human body-aware diffusion model that leverages domain-specific knowledge to enhance performance. Specifically, we employ a pretrained body attention module to guide the diffusion model's focus on the foreground, addressing issues caused by blending between the subject and background. We also demonstrate the value of revisiting the language modality of the diffusion model in restoration tasks by seamlessly incorporating text prompt to improve the quality of surface texture and additional clothing and accessories details. Additionally, we introduce a diffusion sampler tailored for fine-grained human body parts, utilizing local semantic information to rectify limb distortions. Lastly, we collect a comprehensive dataset for benchmarking and advancing the field of human body restoration. Extensive experimental validation showcases the superiority of our approach, both quantitatively and qualitatively, over existing methods.

cs.CV↗

arXiv2024

DIBS: Enhancing Dense Video Captioning with Unlabeled Videos via Pseudo Boundary Enrichment and Online Refinement

We present Dive Into the BoundarieS (DIBS), a novel pretraining framework for dense video captioning (DVC), that elaborates on improving the quality of the generated event captions and their associated pseudo event boundaries from unlabeled videos. By leveraging the capabilities of diverse large language models (LLMs), we generate rich DVC-oriented caption candidates and optimize the corresponding pseudo boundaries under several meticulously designed objectives, considering diversity, event-centricity, temporal ordering, and coherence. Moreover, we further introduce a novel online boundary refinement strategy that iteratively improves the quality of pseudo boundaries during training. Comprehensive experiments have been conducted to examine the effectiveness of the proposed technique components. By leveraging a substantial amount of unlabeled video data, such as HowTo100M, we achieve a remarkable advancement on standard DVC datasets like YouCook2 and ActivityNet. We outperform the previous state-of-the-art Vid2Seq across a majority of metrics, achieving this with just 0.4% of the unlabeled video data used for pre-training by Vid2Seq.

cs.CV↗

arXiv2024

Dynamic Information Dissemination Model Incorporating Non-Adjacent Node Interaction

Describing the dynamics of information dissemination within social networks poses a formidable challenge. Despite multiple endeavors aimed at addressing this issue, only a limited number of studies have effectively replicated and forecasted the evolving course of information dissemination. In this paper, we propose a novel model, DM-NAI, which not only considers the information transfer between adjacent users but also takes into account the information transfer between non-adjacent users to comprehensively depict the information dissemination process. Extensive experiments are conducted on six datasets to predict the information dissemination range and the dissemination trend of the social network. The experimental results demonstrate an average prediction accuracy range of 94.62% to 96.71%, respectively, significantly outperforming state-of-the-art solutions. This finding illustrates that considering information transmission between non-adjacent users helps DM-NAI achieve more accurate information dissemination predictions.

cs.SI↗

arXiv2024

Information Dissemination Model Based on User Attitude and Public Opinion Environment

Modeling the information dissemination process in social networks is a challenging problem. Despite numerous attempts to address this issue, existing studies often assume that user attitudes have only one opportunity to alter during the information dissemination process. Additionally, these studies tend to consider the transformation of user attitudes as solely influenced by a single user, overlooking the dynamic and evolving nature of user attitudes and the impact of the public opinion environment. In this paper, we propose a novel model, UAPE, which considers the influence of the aforementioned factors on the information dissemination process. Specifically, UAPE regards the user's attitude towards the topic as dynamically changing, with the change jointly affected by multiple users simultaneously. Furthermore, the joint influence of multiple users can be considered as the impact of the public opinion environment. Extensive experimental results demonstrate that the model achieves an accuracy range of 91.62% to 94.01%, surpassing the performance of existing research.

cs.SI↗

arXiv2024

Magnetostriction-Driven Muon Localization in an Antiferromagnetic Oxide

Magnetostriction drives a rhombohedral distortion in the cubic rock salt antiferromagnet MnO at the Néel temperature $T_{N}=118$ K. As an unexpected consequence we show that this distortion acts to localize the site of an implanted muon due to the accompanying redistribution of electron density. This lifts the degeneracy between equivalent sites, resulting in a single observed muon precession frequency. Above $T_{N}$, the muon instead becomes delocalized around a network of equivalent sites. Our first-principles simulations based on Hubbard-corrected density-functional theory and molecular dynamics are consistent with our experimental data and help to resolve a long-standing puzzle regarding muon data on MnO, as well as having wider applicability to other magnetic oxides.

cond-mat.str-el↗

arXiv2024

Aleth-NeRF: Illumination Adaptive NeRF with Concealing Field Assumption

The standard Neural Radiance Fields (NeRF) paradigm employs a viewer-centered methodology, entangling the aspects of illumination and material reflectance into emission solely from 3D points. This simplified rendering approach presents challenges in accurately modeling images captured under adverse lighting conditions, such as low light or over-exposure. Motivated by the ancient Greek emission theory that posits visual perception as a result of rays emanating from the eyes, we slightly refine the conventional NeRF framework to train NeRF under challenging light conditions and generate normal-light condition novel views unsupervised. We introduce the concept of a "Concealing Field," which assigns transmittance values to the surrounding air to account for illumination effects. In dark scenarios, we assume that object emissions maintain a standard lighting level but are attenuated as they traverse the air during the rendering process. Concealing Field thus compel NeRF to learn reasonable density and colour estimations for objects even in dimly lit situations. Similarly, the Concealing Field can mitigate over-exposed emissions during the rendering stage. Furthermore, we present a comprehensive multi-view dataset captured under challenging illumination conditions for evaluation. Our code and dataset available at https://github.com/cuiziteng/Aleth-NeRF

cs.CV↗

arXiv2023

Aleth-NeRF: Low-light Condition View Synthesis with Concealing Fields

Common capture low-light scenes are challenging for most computer vision techniques, including Neural Radiance Fields (NeRF). Vanilla NeRF is viewer-centred simplifies the rendering process only as light emission from 3D locations in the viewing direction, thus failing to model the low-illumination induced darkness. Inspired by the emission theory of ancient Greeks that visual perception is accomplished by rays casting from eyes, we make slight modifications on vanilla NeRF to train on multiple views of low-light scenes, we can thus render out the well-lit scene in an unsupervised manner. We introduce a surrogate concept, Concealing Fields, that reduces the transport of light during the volume rendering stage. Specifically, our proposed method, Aleth-NeRF, directly learns from the dark image to understand volumetric object representation and concealing field under priors. By simply eliminating Concealing Fields, we can render a single or multi-view well-lit image(s) and gain superior performance over other 2D low-light enhancement methods. Additionally, we collect the first paired LOw-light and normal-light Multi-view (LOM) datasets for future research. This version is invalid, please refer to our new AAAI version: arXiv:2312.09093

cs.CV↗

arXiv2023

ASM: Adaptive Sample Mining for In-The-Wild Facial Expression Recognition

Given the similarity between facial expression categories, the presence of compound facial expressions, and the subjectivity of annotators, facial expression recognition (FER) datasets often suffer from ambiguity and noisy labels. Ambiguous expressions are challenging to differentiate from expressions with noisy labels, which hurt the robustness of FER models. Furthermore, the difficulty of recognition varies across different expression categories, rendering a uniform approach unfair for all expressions. In this paper, we introduce a novel approach called Adaptive Sample Mining (ASM) to dynamically address ambiguity and noise within each expression category. First, the Adaptive Threshold Learning module generates two thresholds, namely the clean and noisy thresholds, for each category. These thresholds are based on the mean class probabilities at each training epoch. Next, the Sample Mining module partitions the dataset into three subsets: clean, ambiguity, and noise, by comparing the sample confidence with the clean and noisy thresholds. Finally, the Tri-Regularization module employs a mutual learning strategy for the ambiguity subset to enhance discrimination ability, and an unsupervised learning strategy for the noise subset to mitigate the impact of noisy labels. Extensive experiments prove that our method can effectively mine both ambiguity and noise, and outperform SOTA methods on both synthetic noisy and original datasets. The supplement material is available at https://github.com/zzzzzzyang/ASM.

cs.CV↗

arXiv2023

Randomized Quantization: A Generic Augmentation for Data Agnostic Self-supervised Learning

Self-supervised representation learning follows a paradigm of withholding some part of the data and tasking the network to predict it from the remaining part. Among many techniques, data augmentation lies at the core for creating the information gap. Towards this end, masking has emerged as a generic and powerful tool where content is withheld along the sequential dimension, e.g., spatial in images, temporal in audio, and syntactic in language. In this paper, we explore the orthogonal channel dimension for generic data augmentation by exploiting precision redundancy. The data for each channel is quantized through a non-uniform quantizer, with the quantized value sampled randomly within randomly sampled quantization bins. From another perspective, quantization is analogous to channel-wise masking, as it removes the information within each bin, but preserves the information across bins. Our approach significantly surpasses existing generic data augmentation methods, while showing on par performance against modality-specific augmentations. We comprehensively evaluate our approach on vision, audio, 3D point clouds, as well as the DABS benchmark which is comprised of various data modalities. The code is available at https: //github.com/microsoft/random_quantize.

cs.CV↗

arXiv2023

PREFER: Prompt Ensemble Learning via Feedback-Reflect-Refine

As an effective tool for eliciting the power of Large Language Models (LLMs), prompting has recently demonstrated unprecedented abilities across a variety of complex tasks. To further improve the performance, prompt ensemble has attracted substantial interest for tackling the hallucination and instability of LLMs. However, existing methods usually adopt a two-stage paradigm, which requires a pre-prepared set of prompts with substantial manual effort, and is unable to perform directed optimization for different weak learners. In this paper, we propose a simple, universal, and automatic method named PREFER (Pompt Ensemble learning via Feedback-Reflect-Refine) to address the stated limitations. Specifically, given the fact that weak learners are supposed to focus on hard examples during boosting, PREFER builds a feedback mechanism for reflecting on the inadequacies of existing weak learners. Based on this, the LLM is required to automatically synthesize new prompts for iterative refinement. Moreover, to enhance stability of the prompt effect evaluation, we propose a novel prompt bagging method involving forward and backward thinking, which is superior to majority voting and is beneficial for both feedback and weight calculation in boosting. Extensive experiments demonstrate that our PREFER achieves state-of-the-art performance in multiple types of tasks by a significant margin. We have made our code publicly available.

cs.CL↗

arXiv2023

Revisiting Neural Retrieval on Accelerators

Retrieval finds a small number of relevant candidates from a large corpus for information retrieval and recommendation applications. A key component of retrieval is to model (user, item) similarity, which is commonly represented as the dot product of two learned embeddings. This formulation permits efficient inference, commonly known as Maximum Inner Product Search (MIPS). Despite its popularity, dot products cannot capture complex user-item interactions, which are multifaceted and likely high rank. We hence examine non-dot-product retrieval settings on accelerators, and propose \textit{mixture of logits} (MoL), which models (user, item) similarity as an adaptive composition of elementary similarity functions. This new formulation is expressive, capable of modeling high rank (user, item) interactions, and further generalizes to the long tail. When combined with a hierarchical retrieval strategy, \textit{h-indexer}, we are able to scale up MoL to 100M corpus on a single GPU with latency comparable to MIPS baselines. On public datasets, our approach leads to uplifts of up to 77.3\% in hit rate (HR). Experiments on a large recommendation surface at Meta showed strong metric gains and reduced popularity bias, validating the proposed approach's performance and improved generalization.

cs.LG↗

arXiv2023

Long-Term Rhythmic Video Soundtracker

We consider the problem of generating musical soundtracks in sync with rhythmic visual cues. Most existing works rely on pre-defined music representations, leading to the incompetence of generative flexibility and complexity. Other methods directly generating video-conditioned waveforms suffer from limited scenarios, short lengths, and unstable generation quality. To this end, we present Long-Term Rhythmic Video Soundtracker (LORIS), a novel framework to synthesize long-term conditional waveforms. Specifically, our framework consists of a latent conditional diffusion probabilistic model to perform waveform synthesis. Furthermore, a series of context-aware conditioning encoders are proposed to take temporal information into consideration for a long-term generation. Notably, we extend our model's applicability from dances to multiple sports scenarios such as floor exercise and figure skating. To perform comprehensive evaluations, we establish a benchmark for rhythmic video soundtracks including the pre-processed dataset, improved evaluation metrics, and robust generative baselines. Extensive experiments show that our model generates long-term soundtracks with state-of-the-art musical quality and rhythmic correspondence. Codes are available at \url{https://github.com/OpenGVLab/LORIS}.

cs.SD↗