arXiv subjects

Wentao Liu

Publications and source records attributed to Wentao Liu.

At least 73 records · Page 4Linked to original sources

arXiv2024

ShotVL: Human-Centric Highlight Frame Retrieval via Language Queries

Existing works on human-centric video understanding typically focus on analyzing specific moment or entire videos. However, many applications require higher precision at the frame level. In this work, we propose a novel task, BestShot, which aims to locate highlight frames within human-centric videos via language queries. This task demands not only a deep semantic comprehension of human actions but also precise temporal localization. To support this task, we introduce the BestShot Benchmark. %The benchmark is meticulously constructed by combining human detection and tracking, potential frame selection based on human judgment, and detailed textual descriptions crafted by human input to ensure precision. The benchmark is meticulously constructed by combining human-annotated highlight frames, detailed textual descriptions and duration labeling. These descriptions encompass three critical elements: (1) Visual content; (2) Fine-grained action; and (3) Human Pose Description. Together, these elements provide the necessary precision to identify the exact highlight frames in videos. To tackle this problem, we have collected two distinct datasets: (i) ShotGPT4o Dataset, which is algorithmically generated by GPT-4o and (ii) Image-SMPLText Dataset, a dataset with large-scale and accurate per-frame pose description leveraging PoseScript and existing pose estimation datasets. Based on these datasets, we present a strong baseline model, ShotVL, fine-tuned from InternVL, specifically for BestShot. We highlight the impressive zero-shot capabilities of our model and offer comparative analyses with existing SOTA models. ShotVL demonstrates a significant 52% improvement over InternVL on the BestShot Benchmark and a notable 57% improvement on the THUMOS14 Benchmark, all while maintaining the SOTA performance in general image classification and retrieval.

cs.CV↗

arXiv2024

CAS-ViT: Convolutional Additive Self-attention Vision Transformers for Efficient Mobile Applications

Vision Transformers (ViTs) mark a revolutionary advance in neural networks with their token mixer's powerful global context capability. However, the pairwise token affinity and complex matrix operations limit its deployment on resource-constrained scenarios and real-time applications, such as mobile devices, although considerable efforts have been made in previous works. In this paper, we introduce CAS-ViT: Convolutional Additive Self-attention Vision Transformers, to achieve a balance between efficiency and performance in mobile applications. Firstly, we argue that the capability of token mixers to obtain global contextual information hinges on multiple information interactions, such as spatial and channel domains. Subsequently, we propose Convolutional Additive Token Mixer (CATM) employing underlying spatial and channel attention as novel interaction forms. This module eliminates troublesome complex operations such as matrix multiplication and Softmax. We introduce Convolutional Additive Self-attention(CAS) block hybrid architecture and utilize CATM for each block. And further, we build a family of lightweight networks, which can be easily extended to various downstream tasks. Finally, we evaluate CAS-ViT across a variety of vision tasks, including image classification, object detection, instance segmentation, and semantic segmentation. Our M and T model achieves 83.0\%/84.1\% top-1 with only 12M/21M parameters on ImageNet-1K. Meanwhile, throughput evaluations on GPUs, ONNX, and iPhones also demonstrate superior results compared to other state-of-the-art backbones. Extensive experiments demonstrate that our approach achieves a better balance of performance, efficient inference and easy-to-deploy. Our code and model are available at: \url{https://github.com/Tianfang-Zhang/CAS-ViT}

cs.CV↗

arXiv2024

Universal thermodynamic topological classes of rotating black holes

In a recent study, Wei et al. [Phys. Rev. D 110, L081501 (2024)] proposed a universal classification scheme that interprets black hole solutions as topological defects within the thermodynamic parameter space, and then divides black hole solutions into four distinct classes, denoted as $W^{1-}$, $W^{0+}$, $W^{0-}$, and $W^{1+}$, offering insights into deeper aspects of black hole thermodynamics and gravity. In this paper, we investigate the universal thermodynamic topological classification of the singly rotating Kerr black holes in all dimensions, as well as the four-dimensional Kerr-Newman black hole. We show that the innermost small black hole states of the $d \geq 6$ singly rotating Kerr black holes are thermodynamically unstable, while those of the four-dimensional Kerr-Newman black hole and the $d = 4, 5$ singly rotating Kerr black holes are thermodynamically stable. On the other hand, the outermost large black holes exhibit unstable behavior in all these cases. At the low-temperature limit, the $d \geq 6$ singly rotating Kerr black holes have one large thermodynamically unstable black hole, while the four-dimensional Kerr-Newman black hole and the $d = 4, 5$ singly rotating Kerr black holes feature one large unstable branch and one small stable branch. Conversely, at the high-temperature limit, the $d \geq 6$ singly rotating Kerr black holes exhibit a small unstable black hole state, while the four-dimensional Kerr-Newman black hole and the $d = 4, 5$ singly rotating Kerr black holes have no black hole states at all. Consequently, we demonstrate that the $d \geq 6$ singly rotating Kerr black holes belong to the class $W^{1-}$, whereas the four-dimensional Kerr-Newman and $d = 4, 5$ singly rotating Kerr black holes belong to the class $W^{0+}$, thereby further support the conjecture proposed in [Phys. Rev. D 110, L081501 (2024)].

hep-th↗

arXiv2024

KptLLM: Unveiling the Power of Large Language Model for Keypoint Comprehension

Recent advancements in Multimodal Large Language Models (MLLMs) have greatly improved their abilities in image understanding. However, these models often struggle with grasping pixel-level semantic details, e.g., the keypoints of an object. To bridge this gap, we introduce the novel challenge of Semantic Keypoint Comprehension, which aims to comprehend keypoints across different task scenarios, including keypoint semantic understanding, visual prompt-based keypoint detection, and textual prompt-based keypoint detection. Moreover, we introduce KptLLM, a unified multimodal model that utilizes an identify-then-detect strategy to effectively address these challenges. KptLLM underscores the initial discernment of semantics in keypoints, followed by the precise determination of their positions through a chain-of-thought process. With several carefully designed modules, KptLLM adeptly handles various modality inputs, facilitating the interpretation of both semantic contents and keypoint locations. Our extensive experiments demonstrate KptLLM's superiority in various keypoint detection benchmarks and its unique semantic capabilities in interpreting keypoints.

cs.CV↗

arXiv2024

CMM-Math: A Chinese Multimodal Math Dataset To Evaluate and Enhance the Mathematics Reasoning of Large Multimodal Models

Large language models (LLMs) have obtained promising results in mathematical reasoning, which is a foundational skill for human intelligence. Most previous studies focus on improving and measuring the performance of LLMs based on textual math reasoning datasets (e.g., MATH, GSM8K). Recently, a few researchers have released English multimodal math datasets (e.g., MATHVISTA and MATH-V) to evaluate the effectiveness of large multimodal models (LMMs). In this paper, we release a Chinese multimodal math (CMM-Math) dataset, including benchmark and training parts, to evaluate and enhance the mathematical reasoning of LMMs. CMM-Math contains over 28,000 high-quality samples, featuring a variety of problem types (e.g., multiple-choice, fill-in-the-blank, and so on) with detailed solutions across 12 grade levels from elementary to high school in China. Specifically, the visual context may be present in the questions or opinions, which makes this dataset more challenging. Through comprehensive analysis, we discover that state-of-the-art LMMs on the CMM-Math dataset face challenges, emphasizing the necessity for further improvements in LMM development. We also propose a Multimodal Mathematical LMM (Math-LMM) to handle the problems with mixed input of multiple images and text segments. We train our model using three stages, including foundational pre-training, foundational fine-tuning, and mathematical fine-tuning. The extensive experiments indicate that our model effectively improves math reasoning performance by comparing it with the SOTA LMMs over three multimodal mathematical datasets.

cs.CL↗

arXiv2024

PhoCoLens: Photorealistic and Consistent Reconstruction in Lensless Imaging

Lensless cameras offer significant advantages in size, weight, and cost compared to traditional lens-based systems. Without a focusing lens, lensless cameras rely on computational algorithms to recover the scenes from multiplexed measurements. However, current algorithms struggle with inaccurate forward imaging models and insufficient priors to reconstruct high-quality images. To overcome these limitations, we introduce a novel two-stage approach for consistent and photorealistic lensless image reconstruction. The first stage of our approach ensures data consistency by focusing on accurately reconstructing the low-frequency content with a spatially varying deconvolution method that adjusts to changes in the Point Spread Function (PSF) across the camera's field of view. The second stage enhances photorealism by incorporating a generative prior from pre-trained diffusion models. By conditioning on the low-frequency content retrieved in the first stage, the diffusion model effectively reconstructs the high-frequency details that are typically lost in the lensless imaging process, while also maintaining image fidelity. Our method achieves a superior balance between data fidelity and visual quality compared to existing methods, as demonstrated with two popular lensless systems, PhlatCam and DiffuserCam. Project website: https://phocolens.github.io/.

eess.IV↗

arXiv2024

Light rings and shadows of static black holes in effective quantum gravity

Recently, two types of static black hole models that retain general covariance have been proposed within the Hamiltonian constraint approach to effective quantum gravity (EQG). We have studied the light rings and shadows of these black holes using the topological method and the backward ray-tracing method, respectively. We demonstrate that these light rings in both types of static black holes are standard and unstable according to the classification of light rings. Subsequently, we checked the position of the light rings using the photon trajectory equation. We found that although the quantum parameters do not affect the light rings of these two types of black holes, they do reduce the size of the first type of static black hole in EQG, making it smaller. However, for the second type of static black hole in EQG, we cannot distinguish it from a Schwarzschild black hole based on the shadow alone. Fortunately, the quantum parameters shrink the lensing rings of both types of black holes in EQG, causing the black hole shadow to occupy a larger proportion within the ring. This can serve as a basis for distinguishing whether the black hole is in EQG or general relativity (GR).

gr-qc↗

arXiv2024

Static neutral black holes in Kalb-Ramond gravity

The Kalb-Ramond (KR) gravity theory, a modified gravity theory that nonminimally couples a KR field with a nonzero vacuum expectation value for the gravitational field, can spontaneously break the Lorentz symmetry of gravity. In a recent work, Yang et al. [Phys. Rev. D 108, 124004 (2023)] successfully derived Schwarzschild-like black hole solutions both with and without a nonzero cosmological constant within the framework of KR gravity. However, their analysis did not address the more general case of static, neutral, spherically symmetric black holes. In this paper, we fill this gap by resolving the field equations to construct more general static, neutral, spherically symmetric black hole solutions both with and without a nonzero cosmological constant. Our black hole solutions are shown to obey the first law and the Bekenstein-Smarr mass formulas of black hole thermodynamics. Moreover, we demonstrate that our static neutral spherically symmetric AdS black hole does not always satisfy the reverse isoperimetric inequality (RII), as the isoperimetric ratio can be larger or smaller than unity depending on the placement of the solution parameters within the parameter space. This behavior contrasts with the above-mentioned Schwarzschild-like AdS black hole in the KR gravity theory, which always obeys the RII. Significantly, the present more general static, neutral, spherically symmetric AdS black hole is the first example of a static AdS black hole that can violate the RII.

hep-th↗

arXiv2024

GKGNet: Group K-Nearest Neighbor based Graph Convolutional Network for Multi-Label Image Recognition

Multi-Label Image Recognition (MLIR) is a challenging task that aims to predict multiple object labels in a single image while modeling the complex relationships between labels and image regions. Although convolutional neural networks and vision transformers have succeeded in processing images as regular grids of pixels or patches, these representations are sub-optimal for capturing irregular and discontinuous regions of interest. In this work, we present the first fully graph convolutional model, Group K-nearest neighbor based Graph convolutional Network (GKGNet), which models the connections between semantic label embeddings and image patches in a flexible and unified graph structure. To address the scale variance of different objects and to capture information from multiple perspectives, we propose the Group KGCN module for dynamic graph construction and message passing. Our experiments demonstrate that GKGNet achieves state-of-the-art performance with significantly lower computational costs on the challenging multi-label datasets, i.e., MS-COCO and VOC2007 datasets. Codes are available at https://github.com/jin-s13/GKGNet.

cs.CV↗

arXiv2024

UniFS: Universal Few-shot Instance Perception with Point Representations

Instance perception tasks (object detection, instance segmentation, pose estimation, counting) play a key role in industrial applications of visual models. As supervised learning methods suffer from high labeling cost, few-shot learning methods which effectively learn from a limited number of labeled examples are desired. Existing few-shot learning methods primarily focus on a restricted set of tasks, presumably due to the challenges involved in designing a generic model capable of representing diverse tasks in a unified manner. In this paper, we propose UniFS, a universal few-shot instance perception model that unifies a wide range of instance perception tasks by reformulating them into a dynamic point representation learning framework. Additionally, we propose Structure-Aware Point Learning (SAPL) to exploit the higher-order structural relationship among points to further enhance representation learning. Our approach makes minimal assumptions about the tasks, yet it achieves competitive results compared to highly specialized and well optimized specialist models. Codes and data are available at https://github.com/jin-s13/UniFS.

cs.CV↗

arXiv2024

Kerr-MOG-(A)dS black hole and its shadow in scalar-tensor-vector gravity theory

The scalar-tensor-vector gravity (STVG) theory has attracted significant interest due to its ability to effectively address the issue of galaxy rotation curves and clusters of galaxies without considering the influence of dark matter. In this paper, we construct rotating black hole solutions with a cosmological constant in the STVG theory (i.e., Kerr-MOG-(A)dS black hole solutions), where the import of a gravitational charge as a source modifies the gravitational constant, determined by $ G=G_{\text{N}}(1+α) $. For Kerr-MOG-dS spacetime, the observer is situated at a specific location within the domain of outer communication, rather than being located infinitely far away. Since black hole shadows are shaped by light propagation in spacetime, the interaction between the MOG parameter and the cosmological constant is expected to produce novel effects on these shadows. As the cosmological constant $Λ$ increases, the apparent size of the black hole shadow decreases. Additionally, the shadow expands with an increase in the MOG parameter $α$, reaching a maximum at a certain value, and its shape becomes more rounded under an arbitrary rotation parameter, which leads to degeneracy between different black hole parameters. However, by employing numerical ray-tracing techniques, we have found that gravitational lensing and the frame-dragging effect effectively distinguish this degeneracy. Our work contributes to a deeper understanding of black holes in modified gravity, their observational signatures, and constraints.

gr-qc↗

arXiv2024

TCFormer: Visual Recognition via Token Clustering Transformer

Transformers are widely used in computer vision areas and have achieved remarkable success. Most state-of-the-art approaches split images into regular grids and represent each grid region with a vision token. However, fixed token distribution disregards the semantic meaning of different image regions, resulting in sub-optimal performance. To address this issue, we propose the Token Clustering Transformer (TCFormer), which generates dynamic vision tokens based on semantic meaning. Our dynamic tokens possess two crucial characteristics: (1) Representing image regions with similar semantic meanings using the same vision token, even if those regions are not adjacent, and (2) concentrating on regions with valuable details and represent them using fine tokens. Through extensive experimentation across various applications, including image classification, human pose estimation, semantic segmentation, and object detection, we demonstrate the effectiveness of our TCFormer. The code and models for this work are available at https://github.com/zengwang430521/TCFormer.

cs.CV↗

arXiv2024

You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perception

Human-centric perception (e.g. detection, segmentation, pose estimation, and attribute analysis) is a long-standing problem for computer vision. This paper introduces a unified and versatile framework (HQNet) for single-stage multi-person multi-task human-centric perception (HCP). Our approach centers on learning a unified human query representation, denoted as Human Query, which captures intricate instance-level features for individual persons and disentangles complex multi-person scenarios. Although different HCP tasks have been well-studied individually, single-stage multi-task learning of HCP tasks has not been fully exploited in the literature due to the absence of a comprehensive benchmark dataset. To address this gap, we propose COCO-UniHuman benchmark to enable model development and comprehensive evaluation. Experimental results demonstrate the proposed method's state-of-the-art performance among multi-task HCP models and its competitive performance compared to task-specific HCP models. Moreover, our experiments underscore Human Query's adaptability to new HCP tasks, thus demonstrating its robust generalization capability. Codes and data are available at https://github.com/lishuhuai527/COCO-UniHuman.

cs.CV↗

arXiv2024

When Pedestrian Detection Meets Multi-Modal Learning: Generalist Model and Benchmark Dataset

Recent years have witnessed increasing research attention towards pedestrian detection by taking the advantages of different sensor modalities (e.g. RGB, IR, Depth, LiDAR and Event). However, designing a unified generalist model that can effectively process diverse sensor modalities remains a challenge. This paper introduces MMPedestron, a novel generalist model for multimodal perception. Unlike previous specialist models that only process one or a pair of specific modality inputs, MMPedestron is able to process multiple modal inputs and their dynamic combinations. The proposed approach comprises a unified encoder for modal representation and fusion and a general head for pedestrian detection. We introduce two extra learnable tokens, i.e. MAA and MAF, for adaptive multi-modal feature fusion. In addition, we construct the MMPD dataset, the first large-scale benchmark for multi-modal pedestrian detection. This benchmark incorporates existing public datasets and a newly collected dataset called EventPed, covering a wide range of sensor modalities including RGB, IR, Depth, LiDAR, and Event data. With multi-modal joint training, our model achieves state-of-the-art performance on a wide range of pedestrian detection benchmarks, surpassing leading models tailored for specific sensor modality. For example, it achieves 71.1 AP on COCO-Persons and 72.6 AP on LLVIP. Notably, our model achieves comparable performance to the InternImage-H model on CrowdHuman with 30x smaller parameters. Codes and data are available at https://github.com/BubblyYi/MMPedestron.

cs.CV↗

arXiv2024

Shadow of slowly rotating Kalb-Ramond black holes

Real astronomical objects possess spin, yet deriving exact solutions for rotating black holes within gravitational theories is a formidable challenge. To understand the shadow of rotating black holes in Lorentz-violating spacetimes induced by antisymmetric tensor fields, known as Kalb-Ramond (KR) fields, we have focused on the slow-rotation approximation framework. Using this approach, we have obtained first-order rotation series solutions, which describe slowly rotating KR black holes. For this solutions, we have plotted the black hole shadow contours under various parameters using the numerical backward ray-tracing method. As the Lorentz-violating parameter increases, not only the apparent size of the black hole shadow decreases, but also the effects of rotation, such as the D-shaped structure and frame-dragging, are amplified. Furthermore, the KR field also enhances gravitational lensing, causing the shadow to occupy a larger area within the photon ring. This distinctive feature can differentiate KR gravity from general relativity. Additionally, using the latest observational data from EHT on M87* and Sgr A*, we have provided constraints on the Lorentz-violating parameter of rotating KR black holes. We found that, compared to static black holes, rotating black holes allow for the presence of stronger Lorentz violation effects.

gr-qc↗

arXiv2024

DIAS: A Dataset and Benchmark for Intracranial Artery Segmentation in DSA sequences

The automated segmentation of Intracranial Arteries (IA) in Digital Subtraction Angiography (DSA) plays a crucial role in the quantification of vascular morphology, significantly contributing to computer-assisted stroke research and clinical practice. Current research primarily focuses on the segmentation of single-frame DSA using proprietary datasets. However, these methods face challenges due to the inherent limitation of single-frame DSA, which only partially displays vascular contrast, thereby hindering accurate vascular structure representation. In this work, we introduce DIAS, a dataset specifically developed for IA segmentation in DSA sequences. We establish a comprehensive benchmark for evaluating DIAS, covering full, weak, and semi-supervised segmentation methods. Specifically, we propose the vessel sequence segmentation network, in which the sequence feature extraction module effectively captures spatiotemporal representations of intravascular contrast, achieving intracranial artery segmentation in 2D+Time DSA sequences. For weakly-supervised IA segmentation, we propose a novel scribble learning-based image segmentation framework, which, under the guidance of scribble labels, employs cross pseudo-supervision and consistency regularization to improve the performance of the segmentation network. Furthermore, we introduce the random patch-based self-training framework, aimed at alleviating the performance constraints encountered in IA segmentation due to the limited availability of annotated DSA data. Our extensive experiments on the DIAS dataset demonstrate the effectiveness of these methods as potential baselines for future research and clinical applications. The dataset and code are publicly available at https://doi.org/10.5281/zenodo.11396520 and https://github.com/lseventeen/DIAS.

eess.IV↗

arXiv2024

LenslessFace: An End-to-End Optimized Lensless System for Privacy-Preserving Face Verification

Lensless cameras, innovatively replacing traditional lenses for ultra-thin, flat optics, encode light directly onto sensors, producing images that are not immediately recognizable. This compact, lightweight, and cost-effective imaging solution offers inherent privacy advantages, making it attractive for privacy-sensitive applications like face verification. Typical lensless face verification adopts a two-stage process of reconstruction followed by verification, incurring privacy risks from reconstructed faces and high computational costs. This paper presents an end-to-end optimization approach for privacy-preserving face verification directly on encoded lensless captures, ensuring that the entire software pipeline remains encoded with no visible faces as intermediate results. To achieve this, we propose several techniques to address unique challenges from the lensless setup which precludes traditional face detection and alignment. Specifically, we propose a face center alignment scheme, an augmentation curriculum to build robustness against variations, and a knowledge distillation method to smooth optimization and enhance performance. Evaluations under both simulation and real environment demonstrate our method outperforms two-stage lensless verification while enhancing privacy and efficiency. Project website: \url{lenslessface.github.io}.

cs.CV↗

arXiv2024

The RoboDrive Challenge: Drive Anytime Anywhere in Any Condition

In the realm of autonomous driving, robust perception under out-of-distribution conditions is paramount for the safe deployment of vehicles. Challenges such as adverse weather, sensor malfunctions, and environmental unpredictability can severely impact the performance of autonomous systems. The 2024 RoboDrive Challenge was crafted to propel the development of driving perception technologies that can withstand and adapt to these real-world variabilities. Focusing on four pivotal tasks -- BEV detection, map segmentation, semantic occupancy prediction, and multi-view depth estimation -- the competition laid down a gauntlet to innovate and enhance system resilience against typical and atypical disturbances. This year's challenge consisted of five distinct tracks and attracted 140 registered teams from 93 institutes across 11 countries, resulting in nearly one thousand submissions evaluated through our servers. The competition culminated in 15 top-performing solutions, which introduced a range of innovative approaches including advanced data augmentation, multi-sensor fusion, self-supervised learning for error correction, and new algorithmic strategies to enhance sensor robustness. These contributions significantly advanced the state of the art, particularly in handling sensor inconsistencies and environmental variability. Participants, through collaborative efforts, pushed the boundaries of current technologies, showcasing their potential in real-world scenarios. Extensive evaluations and analyses provided insights into the effectiveness of these solutions, highlighting key trends and successful strategies for improving the resilience of driving perception systems. This challenge has set a new benchmark in the field, providing a rich repository of techniques expected to guide future research in this field.

cs.CV↗