Search

Point cloud data now are popular data representations in a number of three-dimensional (3D) vision research realms. However, due to the limited performance of sensors and sensing noise, the raw data usually suffer from sparsity, noise, and incompleteness. This poses great challenges to down-stream point cloud processing tasks. In recent years, deep-learning-based point cloud enhancement methods, which aim to achieve dense, clean, and complete point clouds from low-quality raw point clouds using deep neural networks, are gaining tremendous research attention. This paper, for the first time to our knowledge, presents a comprehensive survey for deep-learning-based point cloud enhancement methods. It covers three main perspectives for point cloud enhancement, i.e., (1) denoising to achieve clean data; (2) completion to recover unseen data; (3) upsampling to obtain dense data. Our survey presents a new taxonomy for recent state-of-the-art methods and systematic experimental results on standard benchmarks. In addition, we share our insightful observations, thoughts, and inspiring future research directions for point cloud enhancement with deep learning.

In the realm of computer vision, the perception and reconstruction of the 3D world through vision signals heavily rely on camera intrinsic parameters, which have long been a subject of intense research within the community. In practical applications, without a strong scene geometry prior like the Manhattan World assumption or special artificial calibration patterns, monocular focal length estimation becomes a challenging task. In this paper, we propose a method for monocular focal length estimation using category-level object priors. Based on two well-studied existing tasks: monocular depth estimation and category-level object canonical representation learning, our focal solver takes depth priors and object shape priors from images containing objects and estimates the focal length from triplets of correspondences in closed form. Our experiments on simulated and real world data demonstrate that the proposed method outperforms the current state-of-the-art, offering a promising solution to the long-standing monocular focal length estimation problem.

Reliable self-localization is a foundational skill for many intelligent mobile platforms. This paper explores the use of event cameras for motion tracking thereby providing a solution with inherent robustness under difficult dynamics and illumination. In order to circumvent the challenge of event camera-based mapping, the solution is framed in a cross-modal way. It tracks a map representation that comes directly from frame-based cameras. Specifically, the proposed method operates on top of gaussian splatting, a state-of-the-art representation that permits highly efficient and realistic novel view synthesis. The key of our approach consists of a novel pose parametrization that uses a reference pose plus first order dynamics for local differential image rendering. The latter is then compared against images of integrated events in a staggered coarse-to-fine optimization scheme. As demonstrated by our results, the realistic view rendering ability of gaussian splatting leads to stable and accurate tracking across a variety of both publicly available and newly recorded data sequences.

Under the backdrop of large-scale pre-training, large visual models (LVM) have demonstrated significant potential in image understanding. The recent emergence of the Segment Anything Model (SAM) has brought a qualitative shift in the field of image segmentation, supporting flexible interactive cues and strong learning capabilities. However, its performance often falls short in cross-domain and few-shot applications. Transferring prior knowledge from foundation models to new applications while preserving learning capabilities is worth exploring. This work proposes a task-adaptive prompt framework based on SAM, a new paradigm for Cross-dominan few-shot segmentation (CD-FSS). First, a Multi-level Feature Fusion (MFF) was used for integrated feature extraction. Besides, an additional Class Domain Task-Adaptive Auto-Prompt (CDTAP) module was combined with the segmentation branch for class-domain agnostic feature extraction and high-quality learnable prompt production. This significant advancement uses a unique generative approach to prompts alongside a comprehensive model structure and specialized prototype computation. While ensuring that the prior knowledge of SAM is not discarded, the new branch disentangles category and domain information through prototypes, guiding it in adapting the CD-FSS. We have achieved the best results on three benchmarks compared to the recent state-of-the-art (SOTA) methods. Comprehensive experiments showed that after task-specific and weighted guidance, the abundant feature information of SAM can be better learned for CD-FSS.

With the introduction of diffusion-based video generation techniques, audio-conditioned human video generation has recently achieved significant breakthroughs in both the naturalness of motion and the synthesis of portrait details. Due to the limited control of audio signals in driving human motion, existing methods often add auxiliary spatial signals to stabilize movements, which may compromise the naturalness and freedom of motion. In this paper, we propose an end-to-end audio-only conditioned video diffusion model named Loopy. Specifically, we designed an inter- and intra-clip temporal module and an audio-to-latents module, enabling the model to leverage long-term motion information from the data to learn natural motion patterns and improving audio-portrait movement correlation. This method removes the need for manually specified spatial motion templates used in existing methods to constrain motion during inference. Extensive experiments show that Loopy outperforms recent audio-driven portrait diffusion models, delivering more lifelike and high-quality results across various scenarios., Comment: Homepage: https://loopyavatar.github.io/

Diffusion-based video generation technology has advanced significantly, catalyzing a proliferation of research in human animation. However, the majority of these studies are confined to same-modality driving settings, with cross-modality human body animation remaining relatively underexplored. In this paper, we introduce, an end-to-end audio-driven human animation framework that ensures hand integrity, identity consistency, and natural motion. The key design of CyberHost is the Region Codebook Attention mechanism, which improves the generation quality of facial and hand animations by integrating fine-grained local features with learned motion pattern priors. Furthermore, we have developed a suite of human-prior-guided training strategies, including body movement map, hand clarity score, pose-aligned reference feature, and local enhancement supervision, to improve synthesis results. To our knowledge, CyberHost is the first end-to-end audio-driven human diffusion model capable of facilitating zero-shot video generation within the scope of human body. Extensive experiments demonstrate that CyberHost surpasses previous works in both quantitative and qualitative aspects., Comment: Homepage: https://cyberhost.github.io/

The high temporal variation of the point clouds is the key challenge of 3D single-object tracking (3D SOT). Existing approaches rely on the assumption that the shape variation of the point clouds and the motion of the objects across neighboring frames are smooth, failing to cope with high temporal variation data. In this paper, we present a novel framework for 3D SOT in point clouds with high temporal variation, called HVTrack. HVTrack proposes three novel components to tackle the challenges in the high temporal variation scenario: 1) A Relative-Pose-Aware Memory module to handle temporal point cloud shape variations; 2) a Base-Expansion Feature Cross-Attention module to deal with similar object distractions in expanded search areas; 3) a Contextual Point Guided Self-Attention module for suppressing heavy background noise. We construct a dataset with high temporal variation (KITTI-HV) by setting different frame intervals for sampling in the KITTI dataset. On the KITTI-HV with 5 frame intervals, our HVTrack surpasses the state-of-the-art tracker CXTracker by 11.3%/15.7% in Success/Precision., Comment: Accepted by ECCV24

Existing neural head avatars methods have achieved significant progress in the image quality and motion range of portrait animation. However, these methods neglect the computational overhead, and to the best of our knowledge, none is designed to run on mobile devices. This paper presents MobilePortrait, a lightweight one-shot neural head avatars method that reduces learning complexity by integrating external knowledge into both the motion modeling and image synthesis, enabling real-time inference on mobile devices. Specifically, we introduce a mixed representation of explicit and implicit keypoints for precise motion modeling and precomputed visual features for enhanced foreground and background synthesis. With these two key designs and using simple U-Nets as backbones, our method achieves state-of-the-art performance with less than one-tenth the computational demand. It has been validated to reach speeds of over 100 FPS on mobile devices and support both video and audio-driven inputs.

Structured light-based method with a camera-projector pair (CPP) plays a vital role in indoor 3D reconstruction, especially for scenes with weak textures. Previous methods usually assume known intrinsics, which are pre-calibrated from known objects, or self-calibrated from multi-view observations. It is still challenging to reliably recover CPP intrinsics from only two views without any known objects. In this paper, we provide a simple yet reliable solution. We demonstrate that, for the first time, sufficient constraints on CPP intrinsics can be derived from an unknown cuboid corner (C2), e.g. a room's corner, which is a common structure in indoor scenes. In addition, with only known camera principal point, the complex multi-variable estimation of all CPP intrinsics can be simplified to a simple univariable optimization problem, leading to reliable calibration and thus direct 3D reconstruction with unknown CPP. Extensive results have demonstrated the superiority of the proposed method over both traditional and learning-based counterparts. Furthermore, the proposed method also demonstrates impressive potential to solve similar tasks without active lighting, such as sparse-view structure from motion.

Foundation models (FMs) are revolutionizing the analysis and understanding of remote sensing (RS) scenes, including aerial RGB, multispectral, and SAR images. However, hyperspectral images (HSIs), which are rich in spectral information, have not seen much application of FMs, with existing methods often restricted to specific tasks and lacking generality. To fill this gap, we introduce HyperSIGMA, a vision transformer-based foundation model for HSI interpretation, scalable to over a billion parameters. To tackle the spectral and spatial redundancy challenges in HSIs, we introduce a novel sparse sampling attention (SSA) mechanism, which effectively promotes the learning of diverse contextual features and serves as the basic block of HyperSIGMA. HyperSIGMA integrates spatial and spectral features using a specially designed spectral enhancement module. In addition, we construct a large-scale hyperspectral dataset, HyperGlobal-450K, for pre-training, which contains about 450K hyperspectral images, significantly surpassing existing datasets in scale. Extensive experiments on various high-level and low-level HSI tasks demonstrate HyperSIGMA's versatility and superior representational capability compared to current state-of-the-art methods. Moreover, HyperSIGMA shows significant advantages in scalability, robustness, cross-modal transferring capability, and real-world applicability., Comment: The code and models will be released at https://github.com/WHU-Sigma/HyperSIGMA

Few-shot semantic segmentation (FSS) endeavors to segment unseen classes with only a few labeled samples. Current FSS methods are commonly built on the assumption that their training and application scenarios share similar domains, and their performances degrade significantly while applied to a distinct domain. To this end, we propose to leverage the cutting-edge foundation model, the Segment Anything Model (SAM), for generalization enhancement. The SAM however performs unsatisfactorily on domains that are distinct from its training data, which primarily comprise natural scene images, and it does not support automatic segmentation of specific semantics due to its interactive prompting mechanism. In our work, we introduce APSeg, a novel auto-prompt network for cross-domain few-shot semantic segmentation (CD-FSS), which is designed to be auto-prompted for guiding cross-domain segmentation. Specifically, we propose a Dual Prototype Anchor Transformation (DPAT) module that fuses pseudo query prototypes extracted based on cycle-consistency with support prototypes, allowing features to be transformed into a more stable domain-agnostic space. Additionally, a Meta Prompt Generator (MPG) module is introduced to automatically generate prompt embeddings, eliminating the need for manual visual prompts. We build an efficient model which can be applied directly to target domains without fine-tuning. Extensive experiments on four cross-domain datasets show that our model outperforms the state-of-the-art CD-FSS method by 5.24% and 3.10% in average accuracy on 1-shot and 5-shot settings, respectively., Comment: 15 pages, 9 figures

In the field of environmental toxicology, rapid and precise assessment of the inflammatory response to pollutants in biological models is critical. This study leverages the power of deep learning to enable automated assessments of zebrafish, a model organism widely used for its translational relevance to human disease pathways. We present an innovative approach to assessing inflammatory responses in zebrafish exposed to various pollutants through an end-to-end deep learning model. The model employs a Unet-based architecture to automatically process high-throughput lateral zebrafish images, segmenting specific regions and quantifying neutrophils as inflammation markers. Alongside imaging, qPCR analysis offers gene expression insights, revealing the molecular impact of exposure on inflammatory pathways. Moreover, the deep learning model was packaged as a user-friendly executable file (.exe), facilitating widespread application by enabling use on virtually any computer without the need for specialized software or training.

Concepts involved in long-form videos such as people, objects, and their interactions, can be viewed as following an implicit prior. They are notably complex and continue to pose challenges to be comprehensively learned. In recent years, generative pre-training (GPT) has exhibited versatile capacities in modeling any kind of text content even visual locations. Can this manner work for learning long-form video prior? Instead of operating on pixel space, it is efficient to employ visual locations like bounding boxes and keypoints to represent key information in videos, which can be simply discretized and then tokenized for consumption by GPT. Due to the scarcity of suitable data, we create a new dataset called \textbf{Storyboard20K} from movies to serve as a representative. It includes synopses, shot-by-shot keyframes, and fine-grained annotations of film sets and characters with consistent IDs, bounding boxes, and whole body keypoints. In this way, long-form videos can be represented by a set of tokens and be learned via generative pre-training. Experimental results validate that our approach has great potential for learning long-form video prior. Code and data will be released at \url{https://github.com/showlab/Long-form-Video-Prior}.

Mycoplasma pneumoniae pneumonia (MPP) poses significant diagnostic challenges in pediatric healthcare, especially in regions like China where it's prevalent. We introduce PneumoniaAPP, a mobile application leveraging deep learning techniques for rapid MPP detection. Our approach capitalizes on convolutional neural networks (CNNs) trained on a comprehensive dataset comprising 3345 chest X-ray (CXR) images, which includes 833 CXR images revealing MPP and additionally augmented with samples from a public dataset. The CNN model achieved an accuracy of 88.20% and an AUROC of 0.9218 across all classes, with a specific accuracy of 97.64% for the mycoplasma class, as demonstrated on the testing dataset. Furthermore, we integrated explainability techniques into PneumoniaAPP to aid respiratory physicians in lung opacity localization. Our contribution extends beyond existing research by targeting pediatric MPP, emphasizing the age group of 0-12 years, and prioritizing deployment on mobile devices. This work signifies a significant advancement in pediatric pneumonia diagnosis, offering a reliable and accessible tool to alleviate diagnostic burdens in healthcare settings., Comment: 27 Pages,7 figures

Talking face generation technology creates talking videos from arbitrary appearance and motion signal, with the "arbitrary" offering ease of use but also introducing challenges in practical applications. Existing methods work well with standard inputs but suffer serious performance degradation with intricate real-world ones. Moreover, efficiency is also an important concern in deployment. To comprehensively address these issues, we introduce SuperFace, a teacher-student framework that balances quality, robustness, cost and editability. We first propose a simple but effective teacher model capable of handling inputs of varying qualities to generate high-quality results. Building on this, we devise an efficient distillation strategy to acquire an identity-specific student model that maintains quality with significantly reduced computational load. Our experiments validate that SuperFace offers a more comprehensive solution than existing methods for the four mentioned objectives, especially in reducing FLOPs by 99\% with the student model. SuperFace can be driven by both video and audio and allows for localized facial attributes editing.

Multi-instance registration is a challenging problem in computer vision and robotics, where multiple instances of an object need to be registered in a standard coordinate system. In this work, we propose the first iterative framework called instance-by-instance (IBI) for multi-instance 3D registration (MI-3DReg). It successively registers all instances in a given scenario, starting from the easiest and progressing to more challenging ones. Throughout the iterative process, outliers are eliminated continuously, leading to an increasing inlier rate for the remaining and more challenging instances. Under the IBI framework, we further propose a sparse-to-dense-correspondence-based multi-instance registration method (IBI-S2DC) to achieve robust MI-3DReg. Experiments on the synthetic and real datasets have demonstrated the effectiveness of IBI and suggested the new state-of-the-art performance of IBI-S2DC, e.g., our MHF1 is 12.02%/12.35% higher than the existing state-of-the-art method ECC on the synthetic/real datasets., Comment: 14 pages, 12 figures, 10 tables

One-shot 3D talking portrait generation aims to reconstruct a 3D avatar from an unseen image, and then animate it with a reference video or audio to generate a talking portrait video. The existing methods fail to simultaneously achieve the goals of accurate 3D avatar reconstruction and stable talking face animation. Besides, while the existing works mainly focus on synthesizing the head part, it is also vital to generate natural torso and background segments to obtain a realistic talking portrait video. To address these limitations, we present Real3D-Potrait, a framework that (1) improves the one-shot 3D reconstruction power with a large image-to-plane model that distills 3D prior knowledge from a 3D face generative model; (2) facilitates accurate motion-conditioned animation with an efficient motion adapter; (3) synthesizes realistic video with natural torso movement and switchable background using a head-torso-background super-resolution model; and (4) supports one-shot audio-driven talking face generation with a generalizable audio-to-motion model. Extensive experiments show that Real3D-Portrait generalizes well to unseen identities and generates more realistic talking portrait videos compared to previous methods. Video samples and source code are available at https://real3dportrait.github.io ., Comment: ICLR 2024 (Spotlight). Project page: https://real3dportrait.github.io

Moir\'e superlattices have emerged as a new platform for studying strongly correlated quantum phenomena, but these systems have been largely limited to van der Waals layer two-dimensional (2D) materials. Here we introduce moir\'e superlattices leveraging ultra-thin, ligand-free halide perovskites, facilitated by ionic interactions. Square moir\'e superlattices with varying periodic lengths are clearly visualized through high-resolution transmission electron microscopy. Twist-angle-dependent transient photoluminescence microscopy and electrical characterizations indicate the emergence of localized bright excitons and trapped charge carriers near a twist angle of ~10{\deg}. The localized excitons are accompanied by enhanced exciton emission, attributed to an increased oscillator strength by a theoretically forecasted flat band. This work illustrates the potential of extended ionic interaction in realizing moir\'e physics at room temperature, broadening the horizon for future investigations.

We present our ongoing work aimed at accelerating a particle-resolved direct numerical simulation model designed to study aerosol-cloud-turbulence interactions. The dynamical model consists of two main components - a set of fluid dynamics equations for air velocity, temperature, and humidity, coupled with a set of equations for particle (i.e., cloud droplet) tracing. Rather than attempting to replace the original numerical solution method in its entirety with a machine learning (ML) method, we consider developing a hybrid approach. We exploit the potential of neural operator learning to yield fast and accurate surrogate models and, in this study, develop such surrogates for the velocity and vorticity fields. We discuss results from numerical experiments designed to assess the performance of ML architectures under consideration as well as their suitability for capturing the behavior of relevant dynamical systems.

Data collected by different modalities can provide a wealth of complementary information, such as hyperspectral image (HSI) to offer rich spectral-spatial properties, synthetic aperture radar (SAR) to provide structural information about the Earth's surface, and light detection and ranging (LiDAR) to cover altitude information about ground elevation. Therefore, a natural idea is to combine multimodal images for refined and accurate land-cover interpretation. Although many efforts have been attempted to achieve multi-source remote sensing image classification, there are still three issues as follows: 1) indiscriminate feature representation without sufficiently considering modal heterogeneity, 2) abundant features and complex computations associated with modeling long-range dependencies, and 3) overfitting phenomenon caused by sparsely labeled samples. To overcome the above barriers, a transformer-based heterogeneously salient graph representation (THSGR) approach is proposed in this paper. First, a multimodal heterogeneous graph encoder is presented to encode distinctively non-Euclidean structural features from heterogeneous data. Then, a self-attention-free multi-convolutional modulator is designed for effective and efficient long-term dependency modeling. Finally, a mean forward is put forward in order to avoid overfitting. Based on the above structures, the proposed model is able to break through modal gaps to obtain differentiated graph representation with competitive time cost, even for a small fraction of training samples. Experiments and analyses on three benchmark datasets with various state-of-the-art (SOTA) methods show the performance of the proposed approach.

Expanding the pool of stable halide perovskites with attractive optoelectronic properties is crucial to addressing current limitations in their performance as photovoltaic (PV) absorbers. In this article, we demonstrate how a high-throughput density functional theory (DFT) dataset of halide perovskite alloys can be used to train accurate surrogate models for property prediction and subsequently perform inverse design using genetic algorithm (GA). Our dataset consists of decomposition energies, band gaps, and photovoltaic efficiencies of nearly 800 pure and mixed composition ABX$_3$ compounds from both the GGA-PBE and HSE06 functionals, and are combined with ~ 100 experimental data points collected from the literature. Multi-fidelity random forest regression models are trained on the DFT + experimental dataset for each property using descriptors that one-hot encode composition, phase, and fidelity, and additionally include well-known elemental or molecular properties of species at the A, B, and X sites. Rigorously optimized models are deployed for experiment-level prediction over > 150,000 hypothetical compounds, leading to thousands of promising materials with low decomposition energy, band gap between 1 and 2 eV, and efficiency > 15%. Surrogate models are further combined with GA using an objective function to maintain chemical feasibility, minimize decomposition energy, maximize PV efficiency, and keep band gap between 1 and 2 eV; hundreds more optimal compositions and phases are thus discovered. We present an analysis of the screened and inverse-designed materials, visualize ternary phase diagrams generated for many systems of interest using ML predictions, and suggest strategies for further improvement and expansion in the future.

Halide perovskites have been extensively studied as materials of interest for optoelectronic applications. There is a major emphasis on ways to tailor the stability, defect behavior, electronic band structure, and optical absorption in halide perovskites, by changing the composition or structure. In this work, we present our contribution to this field in the form of a comprehensive computational investigation of properties as a function of the perovskite phase, different degrees of lattice strains and octahedral distortion and rotation, and the ordering of cations in perovskite alloys. We performed first principles-based density functional theory computations using multiple semi-local and non-local hybrid functionals to calculate optimized lattice parameters, energies of decomposition, electronic band gaps, and theoretical photovoltaic efficiencies. Trends and critical observations from the high-throughput dataset are discussed, especially in terms of the range of optoelectronic properties achievable while keeping the material in a (meta)stable phase or distorted, strained, or differently ordered polymorph. All data is made openly available to the community and is currently being utilized to train state-of-the-art machine learning models for accelerated prediction and discovery, as well as to guide rational experimental discovery.

In these notes we want, in addition to presenting some new results, to both clean up and refine some reflections on a couple of articles published on paper a few years ago, 2019-20. These papers concerned integral sufficient conditions on $u\,,$ $\n u\,,$ and mixed, to guarantee the equality of the energy, EE in the sequel, for solutions of the Navier-Stokes equations under the classical non-slip boundary condition. Concerning the $\n u$ case a crucial role was enjoyed by a previous well known Berselli and Chiodaroli's pioneering 2019 work on the subject. The above three papers are the main sources of these notes. References will be mostly concentrated on their direct relation to the above papers at the time of pubblication. More recent results will be not stated throughout the article. However, in the last section, the reader will be suitably sent to the more recent bibliography. Below, we also turn back to the innovative interpretation of some main parameters which allowed to overcome their apparent incongruence. Non-Newtonian fluids were also considered in our 2019 paper, maybe for the first time in the above Berselli-Chiodaroli's particular $\n u\,$ context. However we will stick mostly to the Newtonian case since in the end we come to the conclusion that there are not particular additional obstacles to extend the present results from Newtonian to non-Newtonian fluids. Hence we avoid to go further in this direction., Comment: 12 pages

Next-generation touch panels are poised to benefit from the use of stretchable and transparent soft ionic conductors, but these materials face several challenges in practical application, including structural damage, loss of functionality, and device stratification, particularly in extreme environments. To address these challenges, in this work, a biocompatible, transparent, self-adhesive gelatin-PAA-based organic hydrogel (PC-OH) was developed, the gel can adhere to the skin in both air and underwater conditions and also anti-freezing, anti-drying, fast self-healable (with a self-healing time of less than 4s in air and underwater), long-term stable for up to 7 days at a wide range of temperatures, highly stretchable, and conductive over a wide temperature range. Using this organic hydrogel, an organic hydrogel-based surface capacitive touch system has been developed that can detect finger touch position in wet environments and over a wide temperature range, demonstrating its ability to sense finger touch position. The wearable touch panel has been successfully demonstrated through the ability to write text, draw figures, and play electronic games, showcasing its potential use in various applications. This breakthrough has significant implications for the development of next-generation touch panels, particularly in the healthcare, sports, and entertainment industries, where reliable and versatile human-machine interfaces are essential., Comment: Unfortunately, there were some disagreements among our research collaborators, which made it impossible for us to proceed with this study. Therefore, we have decided to withdraw this paper so that we can resolve these differences and work together again in the future

Recently there has been a growing interest in category-level object pose and size estimation, and prevailing methods commonly rely on single view RGB-D images. However, one disadvantage of such methods is that they require accurate depth maps which cannot be produced by consumer-grade sensors. Furthermore, many practical real-world situations involve a moving camera that continuously observes its surroundings, and the temporal information of the input video streams is simply overlooked by single-view methods. We propose a novel solution that makes use of RGB video streams. Our framework consists of three modules: a scale-aware monocular dense SLAM solution, a lightweight object pose predictor, and an object-level pose graph optimizer. The SLAM module utilizes a video stream and additional scale-sensitive readings to estimate camera poses and metric depth. The object pose predictor then generates canonical object representations from RGB images. The object pose is estimated through geometric registration of these canonical object representations with estimated object depth points. All per-view estimates finally undergo optimization within a pose graph, culminating in the output of robust and accurate canonical object poses. Our experimental results demonstrate that when utilizing public dataset sequences with high-quality depth information, the proposed method exhibits comparable performance to state-of-the-art RGB-D methods. We also collect and evaluate on new datasets containing depth maps of varying quality to further quantitatively benchmark the proposed method alongside previous RGB-D based methods. We demonstrate a significant advantage in scenarios where depth input is absent or the quality of depth sensing is limited.

Dynamic vision sensors or event cameras provide rich complementary information for video frame interpolation. Existing state-of-the-art methods follow the paradigm of combining both synthesis-based and warping networks. However, few of those methods fully respect the intrinsic characteristics of events streams. Given that event cameras only encode intensity changes and polarity rather than color intensities, estimating optical flow from events is arguably more difficult than from RGB information. We therefore propose to incorporate RGB information in an event-guided optical flow refinement strategy. Moreover, in light of the quasi-continuous nature of the time signals provided by event cameras, we propose a divide-and-conquer strategy in which event-based intermediate frame synthesis happens incrementally in multiple simplified stages rather than in a single, long stage. Extensive experiments on both synthetic and real-world datasets show that these modifications lead to more reliable and realistic intermediate frame results than previous video frame interpolation methods. Our findings underline that a careful consideration of event characteristics such as high temporal density and elevated noise benefits interpolation accuracy., Comment: Accepted by IROS2023 Project Site: https://jiabenchen.github.io/revisit_event