Hongbo Fu

Fudan University · Department of Environmental Science and Engineering

The changes in global air pollutant concentrations influenced by the COVID-19 lockdown have been widely investigated. The lack of clarity regarding the individual contributions to restricted human activities (i.e., transportation) has limited the understanding of the health impacts of the lockdown. In this study, an efficient chemical transport mod...

Personalization techniques for large text‐to‐image (T2I) models allow users to incorporate new concepts from reference images. However, existing methods primarily rely on textual descriptions, leading to limited control over customized images and failing to support fine‐grained and local editing (e.g., shape, pose, and details). In this paper, we i...

Field studies suggest that NO3− photolysis may play a more significant role than previously thought. In this study, we concurrently measured HONO, NO2, and NO2− in situ to gain a deeper understanding of the photogenerated HONO transfer to air and to better constrain the rate constants of NO3− photolysis. The presence of fatty acids (e.g., nonanoic...

Field observations have confirmed that halide ions are widely distributed among aerosols from the marine boundary layer and on the surfaces of ice and snow in polar regions. Consequently, the coexistence of halide ions may play a more significant role in nitrate photolysis than previously thought. In this study, we simultaneously measured HONO, NO2...

High-fidelity online 3D scene reconstruction from monocular videos continues to be challenging, especially for coherent and fine-grained geometry reconstruction. The previous learning-based online 3D reconstruction approaches with neural implicit representations have shown a promising ability for coherent scene reconstruction, but often fail to con...

Geometry- and appearance-controlled full-body human image generation is an interesting but challenging task. Existing solutions are either unconditional or dependent on coarse conditions (e.g., pose, text), thus lacking explicit geometry and appearance control of body and garment. Sketching offers such editing ability and has been adopted in variou...

Modeling 3D avatars benefits various application scenarios such as AR/VR, gaming, and filming. Character faces contribute significant diversity and vividity as a vital component of avatars. However, building 3D character face models usually requires a heavy workload with commercial tools, even for experienced artists. Various existing sketch-based...

The negative effects of air pollution, especially fine particulate matter (PM2.5, particles with an aerodynamic diameter of ≤2.5 μm), on human health, climate, and ecosystems are causing significant concern. Nevertheless, little is known about the contributions of emerging pollutants such as plastic particles to PM2.5 due to the lack of continuous...

Creating fine-retouched portrait images is tedious and time-consuming even for professional artists. There exist automatic retouching methods, but they either suffer from over-smoothing artifacts or lack generalization ability. To address such issues, we present StyleRetoucher, a novel automatic portrait image retouching framework, leveraging Style...

We present a novel freehand sketch beautification method, which takes as input a freely drawn sketch of a man-made object and automatically beautifies it both geometrically and structurally. Beautifying a sketch is challenging because of its highly abstract and heavily diverse drawing manner. Existing methods are usually confined to their limited t...

In the process of product design and digital fabrication, the structural analysis of a designed prototype is a fundamental and essential step. However, such a step is usually invisible or inaccessible to designers at the early sketching phase. This limits the user's ability to consider a shape's physical properties and structural soundness. To brid...

Sketch‐based shape modeling aims to bridge the gap between 2D drawing and 3D modeling by providing an intuitive and accessible approach to create 3D shapes from 2D sketches. However, existing methods still suffer from limitations in reconstruction quality and multi‐view interaction friendliness, hindering their practical application. This paper pro...

How tiny aqueous aerosol droplets significantly accelerate the atmospheric multiphase reaction remains an open question. Here, we show both experimental and theoretical evidence that the strong electric field at the air/water interface of aerosol particles can be an important driving force to contribute to the fast oxidation of SO2-forming sulfate....

Face reenactment and reconstruction benefit various applications in self-media, VR, etc. Recent face reenactment methods use 2D facial landmarks to implicitly retarget facial expressions and poses from driving videos to source images, while they suffer from pose and expression preservation issues for cross-identity scenarios, i.e., when the source...

We present a novel freehand sketch beautification method, which takes as input a freely drawn sketch of a man-made object and automatically beautifies it both geometrically and structurally. Beautifying a sketch is challenging because of its highly abstract and heavily diverse drawing manner. Existing methods are usually confined to the distributio...

In the process of product design and digital fabrication, the structural analysis of a designed prototype is a fundamental and essential step. However, such a step is usually invisible or inaccessible to designers at the early sketching phase. This limits the user's ability to consider a shape's physical properties and structural soundness. To brid...

Atmospheric deposition is an important source for the P biological cycle, especially for P‐limited ecosystems. In the present study, 48 hr size‐segregated aerosol samples were collected at an urban site in Shanghai for one year from May 2019 to June 2020. Total phosphorus (TP) presented a bimodal distribution in four seasons, which was mainly contr...

Large-scale datasets and deep generative models have enabled impressive progress in human face reenactment. Existing solutions for face reenactment have focused on processing real face images through facial landmarks by generative models. Different from real human faces, artistic human faces (e.g., those in paintings, cartoons, etc.) often involve...

Volatile organic compounds (VOCs) have important impacts on air quality, atmospheric chemistry and human health. In order to identify the spatiotemporal variations, sources and ozone (O3) and secondary organic aerosol (SOA) formation potentials of the atmospheric VOCs, a concurrent multi-site observation campaign was performed at the supersites of...

Videos are an accessible form of media for analyzing sports postures and providing feedback to athletes. Existing sport-specific systems embed bespoke human pose attributes and thus can be hard to scale for new attributes, especially for users without programming experiences. Some systems retain scalability by directly showing the differences betwe...

Multiple sketch datasets have been proposed to understand how people draw 3D objects. However, such datasets are often of small scale and cover a small set of objects or categories. In addition, these datasets contain freehand sketches mostly from expert users, making it difficult to compare the drawings by expert and novice users, while such compa...

We propose LiDAL, a novel active learning method for 3D LiDAR semantic segmentation by exploiting inter-frame uncertainty among LiDAR frames. Our core idea is that a well-trained model should generate robust results irrespective of viewpoints for scene scanning and thus the inconsistencies in model predictions across frames provide a very reliable...

Flow velocity, potentially linked to dry deposition velocity and wind speed parameters, show non-marginal significance in the heterogeneous oxidation of SO2 on α-Fe2O3 particles. Our uptake measurement results produce lab-based evidence that an increase in flow velocity can potentially strengthen SO2 uptake capability over α-Fe2O3 by nearly one ord...

The importance of dust heterogeneous oxidation in the removal of atmospheric SO2 and formation of sulfate aerosols is not adequately understood. In this study, the Fe-, Ti-, and Al-bearing components, Na+, Cl-, K+, and Ca2+ of the dust surface, were discovered to be closely associated with the heterogeneous formation of sulfate. Regression models w...

Despite of the extensive research in semiconductor photocatalysis with respect to material and device innovations, much of the fundamental aquatic chemistry of those new materials that governs their environmental hazard and implications remains poorly understood. BiVO4 has long been recognized as a promising visible-light-responsive photocatalyst....

We propose LiDAL, a novel active learning method for 3D LiDAR semantic segmentation by exploiting inter-frame uncertainty among LiDAR frames. Our core idea is that a well-trained model should generate robust results irrespective of viewpoints for scene scanning and thus the inconsistencies in model predictions across frames provide a very reliable...

High-fidelity 3D scene reconstruction from monocular videos continues to be challenging, especially for complete and fine-grained geometry reconstruction. The previous 3D reconstruction approaches with neural implicit representations have shown a promising ability for complete scene reconstruction, while their results are often over-smooth and lack...

Multiple sketch datasets have been proposed to understand how people draw 3D objects. However, such datasets are often of small scale and cover a small set of objects or categories. In addition, these datasets contain freehand sketches mostly from expert users, making it difficult to compare the drawings by expert and novice users, while such compa...

In recent years, sparse voxel-based methods have become the state-of-the-arts for 3D semantic segmentation of indoor scenes, thanks to the powerful 3D CNNs. Nevertheless, being oblivious to the underlying geometry, voxel-based methods suffer from ambiguous features on spatially close objects and struggle with handling complex and irregular geometri...

The spatiotemporal concentration variations, sources, and secondary transformation potential of volatile organic compounds (VOCs) were analyzed at three sampling sites, including the rural (DSL), urban (PD) and industrial (JS) sites, in December 2019, in Shanghai, China. The VOC concentrations in the DSL, PD and JS sites were 13.32 ± 3.33 ppb, 24.7...

Volatile organic compounds (VOCs) have important impacts on air quality, climate, and human health. In order to identify the characteristics, sources and secondary formation potentials of the atmospheric VOCs varying with land use types, a concurrent multi-site observation campaign was performed at the supersites of Shanghai, East China, in the fir...

The importance of dust heterogeneous chemistry in the removal of atmospheric SO2 and formation of sulfate aerosols is not adequately understood. In this study, the Fe, Ti, Al-bearing components, Na+, Cl-, K+, and Ca2+ of the dust surface were discovered to be closely associated with the heterogeneous formation of sulfate. Regression models were the...

Videos are accessible media for analyzing sports postures and providing feedback to athletes. Existing video-based coaching systems often present feedback on the correctness of poses by augmenting videos with visual markers either manually by a coach or automatically by computing key parameters from poses. However, previewing and augmenting videos...

LiDAR and camera are two important sensors for 3D object detection in autonomous driving. Despite the increasing popularity of sensor fusion in this field, the robustness against inferior image conditions, e.g., bad illumination and sensor misalignment, is under-explored. Existing fusion methods are easily affected by such conditions, mainly due to...

In this article, we present GCN-Denoiser, a novel feature-preserving mesh denoising method based on graph convolutional networks ( GCNs ). Unlike previous learning-based mesh denoising methods that exploit handcrafted or voxel-based representations for feature learning, our method explores the structure of a triangular mesh itself and introduces a...

Maintaining global consistency continues to be critical for online 3D indoor scene reconstruction. However, it is still challenging to generate satisfactory 3D reconstruction in terms of global consistency for previous approaches using purely geometric analysis, even with bundle adjustment or loop closure techniques. In this paper, we propose a nov...

Recent deep generative models allow real-time generation of hair images from sketch inputs. Existing solutions often require a user-provided binary mask to specify a target hair shape. This not only costs users extra labor but also fails to capture complicated hair boundaries. Those solutions usually encode hair structures via orientation maps, whi...

Biomass burning has been recognized as an important primary source of atmospheric phosphorus (P), but the measurements of P from biomass burning particles are lacking. In this work, emission factors of different P forms, including total P (TP), total dissolved P (TDP), dissolved inorganic P (DIP) and dissolved organic P (DOP), in emission particles...

Recent deep generative models allow real-time generation of hair images from sketch inputs. Existing solutions often require a user-provided binary mask to specify a target hair shape. This not only costs users extra labor but also fails to capture complicated hair boundaries. Those solutions usually encode hair structures via orientation maps, whi...

In this paper, we present GCN-Denoiser, a novel feature-preserving mesh denoising method based on graph convolutional networks (GCNs). Unlike previous learning-based mesh denoising methods that exploit hand-crafted or voxel-based representations for feature learning, our method explores the structure of a triangular mesh itself and introduces a gra...

Online 3D semantic segmentation, which aims to perform real-time 3D scene reconstruction along with semantic segmentation, is an important but challenging topic. A key challenge is to strike a balance between efficiency and segmentation accuracy. There are very few deep-learning-based solutions to this problem, since the commonly used deep represen...

We introduce SketchGNN , a convolutional graph neural network for semantic segmentation and labeling of freehand vector sketches. We treat an input stroke-based sketch as a graph with nodes representing the sampled points along input strokes and edges encoding the stroke structure information. To predict the per-node labels, our SketchGNN uses grap...

In recent years, sparse voxel-based methods have become the state-of-the-arts for 3D semantic segmentation of indoor scenes, thanks to the powerful 3D CNNs. Nevertheless, being oblivious to the underlying geometry, voxel-based methods suffer from ambiguous features on spatially close objects and struggle with handling complex and irregular geometri...

Particulate nitrite is a critical source of hydroxyl radicals; however, it lacks high-resolution methods due to its low abundance and stability to explore its formation mechanism. In this study, a modified versatile aerosol concentration enrichment system (VACES) coupled with ion chromatography (IC) was used to measure particulate NO2– hourly onlin...

The rapid response to the COVID-19 pandemic led to unprecedented decreases in economic activities, thereby reducing the pollutant emissions. A random forest (RF) model was applied to determine the respective contributions of meteorology and anthropogenic emissions to the changes in air quality. The result suggested that the strict lockdown measures...

High loadings of nitrate (NO3-) in the aerosol over China significantly exacerbate the air quality and pose a great threat to ecosystem safety through dry–wet deposition. Unfortunately, limited ground-level observation data make it challenging to fully reflect the spatial pattern of NO3- levels across China. Until now, long-term monthly particulate...

Humic-like substances (HULIS) are of great interest due to their optical and chemical characteristics. In this study, a total of 180 samples of atmospheric particulate matter (PM) of different sizes were collected from summer 2018 to spring 2019, in order to analyze the size distribution, to investigate the seasonal variation and then to identify t...

Removing outlier correspondences is one of the critical steps for successful feature-based point cloud registration. Despite the increasing popularity of introducing deep learning methods in this field, spatial consistency, which is essentially established by a Euclidean transformation between point clouds, has received almost no individual attenti...

Monolayer-ordered gold nanoring arrays were prepared by ion-sputtering method and used as surface enhanced Raman spectroscopy (SERS) substrates to test the individual atmospheric aerosols particle. Compared to other methods used for testing atmospheric aerosols particles, the collection and subsequent detection in our work is performed directly on...

Heavy loadings of sulfate aerosol trigger haze formation and pose great damage to human health in Taiwan Island. Nevertheless, high-resolution spatiotemporal variation of ambient sulfate across Taiwan Island still remained unknown because of the scarce monitoring sites. Thus, we developed a novel ensemble model named extreme gradient boosting coupl...

Nitrous acid (HONO) plays a significant role in tropospheric chemistry. The emission sources of HONO, however, are poorly characterized, which constraints the predictive capabilities of HONO and the associated environmental influences in models. This study measured the emission factors of HONO (EFHONO) from the agricultural residue (including wheat...

High loadings of nitrate (NO3−) in the aerosol over China significantly exacerbates the air quality and poses a great threaten on ecosystem safety through dry/wet deposition. Unfortunately, limited ground-level observation data makes it challenging to fully reflect the spatial pattern of NO3− level across China. Up to date, the long-term monthly NO...

Water-soluble organic carbon (WSOC), which is closely related to biogenic emissions, is of great importance in the atmosphere for its ubiquitous existence and rich abundance. Levoglucosan, a typical WSOC, is usually considered to be stable and thus used as a tracer of biomass burning. However, we found that levoglucosan can be photo-oxidized on min...

Semantic segmentation and semantic edge detection can be seen as two dual problems with close relationships in computer vision. Despite the fast evolution of learning-based 3D semantic segmentation methods, little attention has been drawn to the learning of 3D semantic edge detectors, even less to a joint learning method for the two tasks. In this...

Wet NH4-N deposition plays a significant role in the ecosystem safety in China, and thus it is highly imperative to estimate the national wet NH4-N deposition flux accurately. In this study, a new methodology named space-time ensemble machine-learning model was first applied to constrain the high-resolution NH4-N deposition fluxes over China based...

Recent research has shown that the complexation of metals-organics plays an important role in atmospheric particulate matter, whose health effects should be taken into account. This work investigates the interactions between catechols (CAs), i.e., 4-nitrocatechol (4NC) and 4-methylcatechol (4MC), and transition metals (i.e., Fe) in the aqueous phas...

In this paper, we study the problem of multi-view sketch correspondence, where we take as input multiple freehand sketches with different views of the same object and predict as output the semantic correspondence among the sketches. This problem is challenging since the visual features of corresponding points at different views can be very differen...

Semantic segmentation and semantic edge detection can be seen as two dual problems with close relationships in computer vision. Despite the fast evolution of learning-based 3D semantic segmentation methods, little attention has been drawn to the learning of 3D semantic edge detectors, even less to a joint learning method for the two tasks. In this...

We developed a two-stage model called the random-forest–generalised additive model (RF–GAM), based on satellite data, meteorological factors, and other geographical covariates, to predict the surface 8 h O3 concentrations across the remote Tibetan Plateau. The 10-fold cross-validation result suggested that RF–GAM showed excellent performance, with...

Sketches in existing large-scale datasets like the recent QuickDraw collection are often stored in a vector format, with strokes consisting of sequentially sampled points. However, most existing sketch recognition methods rasterize vector sketches as binary images and then adopt image classification techniques. In this paper, we propose a novel end...

The heterogeneous reaction of SO2 on mineral dust surfaces is generally considered as an important chemical pathway for secondary sulfate formation in the troposphere. To this day, there are no reported studies that assess the impact of atmospheric CO2 in sulfate production on mineral dust surfaces. In this work, we investigate the impact of CO2 on...

We introduce SketchGCN, a graph convolutional neural network for semantic segmentation and labeling of free-hand sketches. We treat an input sketch as a 2D pointset, and encode the stroke structure information into graph node/edge representations. To predict the per-point labels, our SketchGCN uses graph convolution and a global-local branching net...

As a much finer particle, particulate matter less than 1 μm (PM1) plays an important role on the haze formation and human health. However, the capability of mapping PM1 concentration is severely impaired by coarse temporal resolution and low estimation accuracy, largely due to the neglect of spatial or temporal autocorrelation of PM1. In order to i...

China has experienced severe acid rain pollution during the past decades due to excessive sulfur oxides (SO2) and nitrous oxides (NOx) emissions, which further caused lake acidification, biodiversity losses and climate change. Although the major sources of acid deposition have been clarified previously, the contributions of socioeconomic (natural)...