Search

Wenze Han,1,2 Ran Zhang,1,2 Shuzhi Liu,1,2 Tong Zhang,1,2 Xuemin Yao,1,2 Yuxin Cao,1,2 Jiadi Li,1,2 Xiaoming Liu,1,2 Bing Li1,2 1Shanxi Medical University School and Hospital of Stomatology, Taiyuan, Shanxi, 030001, People’s Republic of China; 2Shanxi Province Key Laboratory of Oral Diseases Prevention and New Materials, Taiyuan, 030001, Shanxi, People’s Republic of ChinaCorrespondence: Xiaoming Liu; Bing Li, Shanxi Medical University School and Hospital of Stomatology, Taiyuan, Shanxi, 030001, People’s Republic of China, Email lxming0912@163.com; libing1975vip@163.comAbstract: Whiskers are nanoscale, high-strength fibrous crystals with a wide range of potential applications in dentistry owing to their unique mechanical, thermal, electrical, and biological properties. They possess high strength, a high modulus of elasticity and good biocompatibility. Hence, adding these crystals to dental composites as reinforcement can considerably improve the mechanical properties and durability of restorations. Additionally, whiskers are involved in inducing the value-added differentiation of osteoblasts, odontogenic osteocytes, and pulp stem cells, and promoting the regeneration of alveolar bone, periodontal tissue, and pulp tissue. They can also enhance the mucosal barrier function, inhibit the proliferation of tumor cells, control inflammation, and aid in cancer prevention. This review comprehensively summarizes the classification, properties, growth mechanisms and preparation methods of whiskers and focuses on their application in dentistry. Due to their unique physicochemical properties, excellent biological properties, and nanoscale characteristics, whiskers show great potential for application in bone, periodontal, and pulp tissue regeneration. Additionally, they can be used to prevent and treat oral cancer and improve medical devices, thus making them a promising new material in dentistry.Keywords: whiskers, dentistry, growth mechanism, biomaterials

Xiaoming Liu,1 Xiaoni Yang2 1Department of Geriatric Medicine, The Affiliated Hospital of Shandong University of Traditional Chinese Medicine, Jinan, 250014, People’s Republic of China; 2Department of Traditional Chinese Medicine, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Jinan, 250014, People’s Republic of ChinaCorrespondence: Xiaoni Yang, Department of Traditional Chinese Medicine, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, No. 16766 Jingshi Road, Jinan, 250014, People’s Republic of China, Email yangxiaoni1979@163.comAbstract: Global aging is rapidly accelerating, which significantly influences the health systems worldwide. Frailty emerges as the most conspicuous hallmark of aging, imposing novel global health challenges. Characterized by a multifaceted decline across physiological system, frailty diminishes an individual’s capacity to maintain equilibrium in the presence of stressors, which leads to adverse outcomes such as falls, delirium, and disability. Several screening tools and interventions have been developed to mitigate the harm caused by frailty to human health, but research on frailty in mainland China commences belatedly with scant studies conducted. Therefore, it is imperative to explore screening methods and treatment modalities tailored to the Chinese context, thereby enhancing the older adults’ quality of life and advancing social medicine. This review aims to elucidate the evolution, diagnosis, and management of frailty, alongside the challenges it poses, with the overarching goal of guiding future diagnostic and therapeutic endeavors. Specifically, we summarized the mechanisms of frailty and intervention strategies in elderly people, and meanwhile, we evaluated the advantages and disadvantages of different measurement tools.Keywords: frailty, geriatrics, pathogenesis, assessment, intervention

Two limiting factors of microbial electrochemical denitrification (MED) are the abundance and efficiency of the functional microorganisms. To supply these microorganisms, MED systems are inoculated with denitrifying sludge, but such method has much room for improvement. This study compared MED inoculated with autotrophic denitrifying inoculum (ADI) versus with heterotrophic denitrifying inoculum (HDI). ADI exhibited electroactivity for 50% less of time than HDI. The denitrification efficiency of the ADI biocathode was 42% higher than that of the HDI biocathode. The HDI biocathode had high levels of polysaccharides while the ADI biocathode was rich in proteins, suggesting that two biocathodes may achieve MED but via different pathways. Microbial communities of two biocathodes indicated MED of HDI biocathode may rely on interspecies electron transfer, whereas sulfur bacteria of ADI biocathode take electrons directly from the cathode to achieve MED. Utilizing autotrophic sulfur-oxidizing denitrifiers, this study offers a strategy for enhancing MED. © 2023 Elsevier Ltd

Owing to the relentless growth of sewage sludge production, achieving low-carbon development in sewage sludge treatment and disposal (STD) is becoming increasingly challenging and unpredictable. However, the STD varied spatially, and city-specific analysis is deemed necessary for sustainable evaluation. Therefore, a lifecycle-based greenhouse gas (GHG), energy, and economic analysis were conducted by considering six local STD alternatives in Wuhan City, China, as a case study. The findings indicated anaerobic digestion combined with digestate utilization for urban greening (ADL) and incineration in existing power plants (INCP) exhibited the least GHG emissions at 34.073 kg CO2 eq/FU and 644.128 kg CO2 eq/FU, while INCP generated the most energy at −2594 kW.h/FU. The economic evaluation revealed that ADL and INCP were more beneficial without accounting for land acquisition. Scenario analysis showed that the energy recovery from ADL and INCP is significantly influenced by the hydrolysis yielding rate and sludge organic content. Perturbation sensitivity indicates that regional emission factor of electricity and electricity fee highly influence the overall GHG emission and cost. The results of this study could assist policymakers in identifying viable solutions to the cities experiencing the same sludge treatment burdens. © 2023

Yuping Zhu,1 Xunfu Lai,1 Mengting Wang,2 Xin Tang,1 Tianyi Wan,3 Bin Li,1 Xiaoming Liu,1 Jialin Wu,1 Lei He,1 Yulin He1 1Department of Radiology, The First Affiliated Hospital of Nanchang University, Nanchang, People’s Republic of China; 2Department of Radiology, Yichang Central People’s Hospital, Yichang, People’s Republic of China; 3Department of Radiology, Jiangxi Provincial People’s Hospital, Nanchang, People’s Republic of ChinaCorrespondence: Yulin He, Department of Radiology, The First Affiliated Hospital of Nanchang University, 17 Yongwaizheng Street, Nanchang, Jiangxi, 330006, People’s Republic of China, Tel +86 0791 8869 3802, Email ndyfy02171@ncu.edu.cnObjective: To separate the resting-state network of patients with dental pain using independent component analysis (ICA) and analyze abnormal changes in functional connectivity within as well as between the networks.Patients and Methods: Twenty-three patients with dental pain and 30 healthy controls participated in this study. We extracted the resting-state functional network components of both using ICA. Functional connectivity differences within 14 resting-state brain networks were analyzed at the voxel level. Directional interactions between networks were analyzed using Granger causality analysis. Subsequently, functional connectivity values and causal coefficients were assessed for correlations with clinical parameters.Results: Compared to healthy controls, we found enhanced functional connectivity in the left superior temporal gyrus of anterior protrusion network and the right Rolandic operculum of auditory network in patients with dental pain (p< 0.01 and cluster-level p< 0.05, Gaussian random field corrected). In contrast, functional connectivity of the right precuneus in the precuneus network was reduced, and were significantly as well as negatively correlated to those of the Visual Analogue Scale (r=− 4.93, p=0.017), Hamilton Anxiety Scale (r=− 0.46, p=0.027), and Hamilton Depr

Monocular 3D detectors achieve remarkable performance on cars and smaller objects. However, their performance drops on larger objects, leading to fatal accidents. Some attribute the failures to training data scarcity or their receptive field requirements of large objects. In this paper, we highlight this understudied problem of generalization to large objects. We find that modern frontal detectors struggle to generalize to large objects even on nearly balanced datasets. We argue that the cause of failure is the sensitivity of depth regression losses to noise of larger objects. To bridge this gap, we comprehensively investigate regression and dice losses, examining their robustness under varying error levels and object sizes. We mathematically prove that the dice loss leads to superior noise-robustness and model convergence for large objects compared to regression losses for a simplified case. Leveraging our theoretical insights, we propose SeaBird (Segmentation in Bird's View) as the first step towards generalizing to large objects. SeaBird effectively integrates BEV segmentation on foreground objects for 3D detection, with the segmentation head trained with the dice loss. SeaBird achieves SoTA results on the KITTI-360 leaderboard and improves existing detectors on the nuScenes leaderboard, particularly for large objects. Code and models at https://github.com/abhi1kumar/SeaBird, Comment: CVPR 2024

In this paper, we address the challenge of making ViT models more robust to unseen affine transformations. Such robustness becomes useful in various recognition tasks such as face recognition when image alignment failures occur. We propose a novel method called KP-RPE, which leverages key points (e.g.~facial landmarks) to make ViT more resilient to scale, translation, and pose variations. We begin with the observation that Relative Position Encoding (RPE) is a good way to bring affine transform generalization to ViTs. RPE, however, can only inject the model with prior knowledge that nearby pixels are more important than far pixels. Keypoint RPE (KP-RPE) is an extension of this principle, where the significance of pixels is not solely dictated by their proximity but also by their relative positions to specific keypoints within the image. By anchoring the significance of pixels around keypoints, the model can more effectively retain spatial relationships, even when those relationships are disrupted by affine transformations. We show the merit of KP-RPE in face and gait recognition. The experimental results demonstrate the effectiveness in improving face recognition performance from low-quality images, particularly where alignment is prone to failure. Code and pre-trained models are available., Comment: To appear in CVPR2024

Generative AI (GenAI) is transforming creative workflows through the capability to synthesize and manipulate images via high-level prompts. Yet creatives are not well supported to receive recognition or reward for the use of their content in GenAI training. To this end, we propose ProMark, a causal attribution technique to attribute a synthetically generated image to its training data concepts like objects, motifs, templates, artists, or styles. The concept information is proactively embedded into the input training images using imperceptible watermarks, and the diffusion models (unconditional or conditional) are trained to retain the corresponding watermarks in generated images. We show that we can embed as many as $2^{16}$ unique watermarks into the training data, and each training image can contain more than one watermark. ProMark can maintain image quality whilst outperforming correlation-based attribution. Finally, several qualitative examples are presented, providing the confidence that the presence of the watermark conveys a causative relationship between training data and synthetic images., Comment: Accepted to CVPR 2024

Gait recognition stands as one of the most pivotal remote identification technologies and progressively expands across research and industry communities. However, existing gait recognition methods heavily rely on task-specific upstream driven by supervised learning to provide explicit gait representations like silhouette sequences, which inevitably introduce expensive annotation costs and potential error accumulation. Escaping from this trend, this work explores effective gait representations based on the all-purpose knowledge produced by task-agnostic Large Vision Models (LVMs) and proposes a simple yet efficient gait framework, termed BigGait. Specifically, the Gait Representation Extractor (GRE) within BigGait draws upon design principles from established gait representations, effectively transforming all-purpose knowledge into implicit gait representations without requiring third-party supervision signals. Experiments on CCPG, CAISA-B* and SUSTech1K indicate that BigGait significantly outperforms the previous methods in both within-domain and cross-domain tasks in most cases, and provides a more practical paradigm for learning the next-generation gait representation. Finally, we delve into prospective challenges and promising directions in LVMs-based gait recognition, aiming to inspire future work in this emerging topic. The source code is available at https://github.com/ShiqiYu/OpenGait.

The widespread use of large language models (LLMs) is increasing the demand for methods that detect machine-generated text to prevent misuse. The goal of our study is to stress test the detectors' robustness to malicious attacks under realistic scenarios. We comprehensively study the robustness of popular machine-generated text detectors under attacks from diverse categories: editing, paraphrasing, prompting, and co-generating. Our attacks assume limited access to the generator LLMs, and we compare the performance of detectors on different attacks under different budget levels. Our experiments reveal that almost none of the existing detectors remain robust under all the attacks, and all detectors exhibit different loopholes. Averaging all detectors, the performance drops by 35% across all attacks. Further, we investigate the reasons behind these defects and propose initial out-of-the-box patches to improve robustness.

The rapid advancement of diffusion models (DMs) has not only transformed various real-world industries but has also introduced negative societal concerns, including the generation of harmful content, copyright disputes, and the rise of stereotypes and biases. To mitigate these issues, machine unlearning (MU) has emerged as a potential solution, demonstrating its ability to remove undesired generative capabilities of DMs in various applications. However, by examining existing MU evaluation methods, we uncover several key challenges that can result in incomplete, inaccurate, or biased evaluations for MU in DMs. To address them, we enhance the evaluation metrics for MU, including the introduction of an often-overlooked retainability measurement for DMs post-unlearning. Additionally, we introduce UnlearnCanvas, a comprehensive high-resolution stylized image dataset that facilitates us to evaluate the unlearning of artistic painting styles in conjunction with associated image objects. We show that this dataset plays a pivotal role in establishing a standardized and automated evaluation framework for MU techniques on DMs, featuring 7 quantitative metrics to address various aspects of unlearning effectiveness. Through extensive experiments, we benchmark 5 state-of-the-art MU methods, revealing novel insights into their pros and cons, and the underlying unlearning mechanisms. Furthermore, we demonstrate the potential of UnlearnCanvas to benchmark other generative modeling tasks, such as style transfer. The UnlearnCanvas dataset, benchmark, and the codes to reproduce all the results in this work can be found at https://github.com/OPTML-Group/UnlearnCanvas.

Face Recognition (FR) systems can suffer from physical (i.e., print photo) and digital (i.e., DeepFake) attacks. However, previous related work rarely considers both situations at the same time. This implies the deployment of multiple models and thus more computational burden. The main reasons for this lack of an integrated model are caused by two factors: (1) The lack of a dataset including both physical and digital attacks with ID consistency which means the same ID covers the real face and all attack types; (2) Given the large intra-class variance between these two attacks, it is difficult to learn a compact feature space to detect both attacks simultaneously. To address these issues, we collect a Unified physical-digital Attack dataset, called UniAttackData. The dataset consists of $1,800$ participations of 2 and 12 physical and digital attacks, respectively, resulting in a total of 29,706 videos. Then, we propose a Unified Attack Detection framework based on Vision-Language Models (VLMs), namely UniAttackDetection, which includes three main modules: the Teacher-Student Prompts (TSP) module, focused on acquiring unified and specific knowledge respectively; the Unified Knowledge Mining (UKM) module, designed to capture a comprehensive feature space; and the Sample-Level Prompt Interaction (SLPI) module, aimed at grasping sample-level semantics. These three modules seamlessly form a robust unified attack detection framework. Extensive experiments on UniAttackData and three other datasets demonstrate the superiority of our approach for unified face attack detection., Comment: 12 pages, 8 figures

The burgeoning generative capabilities of large language models (LLMs) have raised growing concerns about abuse, demanding automatic machine-generated text detectors. DetectGPT, a zero-shot metric-based detector, first introduces perturbation and shows great performance improvement. However, in DetectGPT, random perturbation strategy could introduce noise, and logit regression depends on threshold, harming the generalizability and applicability of individual or small-batch inputs. Hence, we propose a novel fine-tuned detector, Pecola, bridging metric-based and fine-tuned detectors by contrastive learning on selective perturbation. Selective strategy retains important tokens during perturbation and weights for multi-pair contrastive learning. The experiments show that Pecola outperforms the state-of-the-art by 1.20% in accuracy on average on four public datasets. And we further analyze the effectiveness, robustness, and generalization of the method.

Synthetic data is gaining increasing relevance for training machine learning models. This is mainly motivated due to several factors such as the lack of real data and intra-class variability, time and errors produced in manual labeling, and in some cases privacy concerns, among others. This paper presents an overview of the 2nd edition of the Face Recognition Challenge in the Era of Synthetic Data (FRCSyn) organized at CVPR 2024. FRCSyn aims to investigate the use of synthetic data in face recognition to address current technological limitations, including data privacy concerns, demographic biases, generalization to novel scenarios, and performance constraints in challenging situations such as aging, pose variations, and occlusions. Unlike the 1st edition, in which synthetic data from DCFace and GANDiffFace methods was only allowed to train face recognition systems, in this 2nd edition we propose new sub-tasks that allow participants to explore novel face generative methods. The outcomes of the 2nd FRCSyn Challenge, along with the proposed experimental protocol and benchmarking contribute significantly to the application of synthetic data to face recognition., Comment: arXiv admin note: text overlap with arXiv:2311.10476

Large language models have shown their ability to become effective few-shot learners with prompting, revoluting the paradigm of learning with data scarcity. However, this approach largely depends on the quality of prompt initialization, and always exhibits large variability among different runs. Such property makes prompt tuning highly unreliable and vulnerable to poorly constructed prompts, which limits its extension to more real-world applications. To tackle this issue, we propose to treat the hard prompt and soft prompt as separate inputs to mitigate noise brought by the prompt initialization. Furthermore, we optimize soft prompts with contrastive learning for utilizing class-aware information in the training process to maintain model performance. Experimental results demonstrate that \sysname outperforms state-of-the-art methods by 7.20% in accuracy and reduces the standard deviation by 2.02 on average. Furthermore, extensive experiments underscore its robustness and stability across 7 datasets covering various tasks., Comment: Submitted to ACL 2024

Food waste (FW) increases with urbanization and population growth, which puts pressure on the treatment system, causing a variety of harmful impacts on the environment. Proper FW treatment is imperative for ecological integrity and public health. Even though FW treatment is an extensively studied topic, the sustainable FW treatment considering holistic-lifecycle-based environmental impacts has rarely been evaluated. This study addresses this gap through a comprehensive analysis of various FW treatment methods, including co-treatment with sewage, anaerobic digestion, incineration, and aerobic composting. The impacts of greenhouse gas (GHG) emission and energy use efficiency are assessed by analyzing diverse FW treatment methods in Shenzhen, China. The study indicates that FW addition to sewage does not adversely affect the current sewage treatment plant, but benefits GHG avoidance and energy recovery. Compared with the other FW treatment methods, FW anaerobic digestion avoids the most GHG emissions with −71.3 kg CO2 eq/FU and recovers the most energy with −223 kWh/FU, followed by FW co-treated with sewage. The energy conversion efficiency of the combined heat and power (CHP) unit greatly affects FW incineration, while energy consumption in incineration and anaerobic digestion (AD) process is relatively minor. Perturbation analysis pinpoints key parameters influencing outcomes, including CHP efficiency, GHG emission factor of local electricity, and chemical oxygen demand (COD) in FW with ratios of −13~−0.942, −0.518~0.22, and −13~1.01, respectively, that should be given special attention. This study sheds light on sustainable FW management strategies, not only in China but also transferrable to regions confronting similar challenges. Advocating ecologically balanced and resource-efficient approaches, the study aligns with broader aims of fostering sustainable development. © 2023 by the authors.

The generation of medical waste (MW) is expected to increase significantly during a public health crisis, such as COVID-19, which presents serious challenges to subsequent MW treatment systems. However, little research has been investigated into the sustainability of MW management during emergency crisis. It is crucial to explore the variations brought about by the pandemic and identify proper measures to improve inadequate treatment efficiencies to handle the increased MW burdens, but, this has not been addressed yet. This study developed an integrated model of life cycle assessment (LCA) with data envelopment analysis (DEA) to examine and enhance the efficiency of MW management during a pandemic, with a focus on Wuhan City, Hubei Province, and China. The findings showed that during the peak of the pandemic, environmental and economic impacts of MW treatment were 1.3–5.4 times and 1.9–5.9 times higher, respectively. These variations were mostly caused by steam sterilization to create a high-temperature environment, emphasizing the need to replace conventional fuels with sustainable sources. The LCA-DEA model identified movable organic Rankine cycle (ORC) incinerator as the most efficient option, achieving a full score of 1.0 for both pure technical efficiency and scale efficiency. Additionally, the model quantified potential reductions in environmental impacts, such as global warming and terrestrial acidification, to enhance management efficiency of inadequate strategies. The findings of this study not only reveal variations brought about by the pandemic but also provide effective measures to enhance MW management efficiencies under emergency situations, which can indicate policymakers and practitioners on how to improve MW management during public health crisis. © 2023

Food waste (FW) being a major solid waste component and of degradable nature is the most challenging to manage and mitigate greenhouse gas emissions (GHEs). Policymakers seek innovative approaches to achieve net zero objectives and recover resources from the FW which requires a comparative and holistic investigation of contemporary treatment methods. This study assessed the lifecycle of six alternative scenarios for reducing net GHEs and energy use potential from FW management in a metropolis, taking Hong Kong as a reference. In both impact categories, the business-as-usual (landfilling) was the worst-case scenario. The combined anaerobic digestion and composting (ADC) technique was ranked best in the global warming impact but was more energy intensive than anaerobic digestion with sludge landfilling (ADL). Incineration ranked second in net GHEs but less favourable for energy recovery from FW alone. The proposed integration of FW and biological wastewater treatment represented an enticing alternative. Integration by co-disposal and treatment with wastewater (CoDT-WW) performed above average in both categories, while anaerobic co-digestion with sewage sludge (AnCoD-SS) ranked fourth. The sensitivity analysis further identified critical parameters inherent to individual scenarios along with biogenic carbon emission and sequestration, revealing their significance on the magnitude of GHEs and scenarios’ ranking. Capacity assessment of the studied treatment facilities showed a FW diversion potential of ∼60% while reducing the net GHEs by ∼70% compared to the base-case, indicating potential of net zero carbon emissions and energy footprint by increasing treatment capacity. From this study, policymakers can gain insights and guidelines for low-carbon urban infrastructure development worldwide. © 2023

Conductive materials have been reported to enhance anaerobic digestion (AD) processes, but their role in the synergistic relationship between syntrophic bacteria and hydrogenotrophic methanogens (HMs) remains unclear yet. This study evaluated the holistic impact of conductive materials in an HMs-dominant anaerobic system (HMs over 95% in the archaeal community). Biochar (BC) and iron powder (IP) were selected as conductive materials with food waste and waste-activated sludge as co-substrates. The addition of IP and BC to HMs-dominant reactors significantly increased the methane yield (11.5–26%) and shortened the digestion period (32–45%) compared with the control. In addition, a novel developed integrated kinetic model revealed the diauxic peak for methanogenesis which could be significantly alleviated by the mediation of conductive materials. Furthermore, the degradation of volatile fatty acids followed a sequential order of acetate–butyrate–propionate. A smoother degradation was observed with the addition of conductive materials, while the control groups exhibited a 10 day lag phase between the degradation of butyrate and propionate. The microbial community analysis showed that BC and IP stimulated diverse syntrophic bacteria, that is, BC-enriched genus Longilinea and Aminicenantales and IP-accumulated Syntrophomonadaceae. Moreover, IP significantly facilitated HMs enrichment. Overall, this study offers new insights into the regulation of conductive materials on microbial pathways of syntrophic and hydrogenotrophic methanogenesis, with potential implications for the development of sustainable and efficient organic waste treatment strategies.

Numerous adversarial attack methods have been developed to generate imperceptible image perturbations that can cause erroneous predictions of state-of-the-art machine learning (ML) models, in particular, deep neural networks (DNNs). Despite intense research on adversarial attacks, little effort was made to uncover 'arcana' carried in adversarial attacks. In this work, we ask whether it is possible to infer data-agnostic victim model (VM) information (i.e., characteristics of the ML model or DNN used to generate adversarial attacks) from data-specific adversarial instances. We call this 'model parsing of adversarial attacks' - a task to uncover 'arcana' in terms of the concealed VM information in attacks. We approach model parsing via supervised learning, which correctly assigns classes of VM's model attributes (in terms of architecture type, kernel size, activation function, and weight sparsity) to an attack instance generated from this VM. We collect a dataset of adversarial attacks across 7 attack types generated from 135 victim models (configured by 5 architecture types, 3 kernel size setups, 3 activation function types, and 3 weight sparsity ratios). We show that a simple, supervised model parsing network (MPN) is able to infer VM attributes from unseen adversarial attacks if their attack settings are consistent with the training setting (i.e., in-distribution generalization assessment). We also provide extensive experiments to justify the feasibility of VM parsing from adversarial attacks, and the influence of training and evaluation factors in the parsing performance (e.g., generalization challenge raised in out-of-distribution evaluation). We further demonstrate how the proposed MPN can be used to uncover the source VM attributes from transfer attacks, and shed light on a potential connection between model parsing and attack transferability.

Recent advances in generative adversarial networks (GANs) have led to remarkable achievements in face image synthesis. While methods that use style-based GANs can generate strikingly photorealistic face images, it is often difficult to control the characteristics of the generated faces in a meaningful and disentangled way. Prior approaches aim to achieve such semantic control and disentanglement within the latent space of a previously trained GAN. In contrast, we propose a framework that a priori models physical attributes of the face such as 3D shape, albedo, pose, and lighting explicitly, thus providing disentanglement by design. Our method, MOST-GAN, integrates the expressive power and photorealism of style-based GANs with the physical disentanglement and flexibility of nonlinear 3D morphable models, which we couple with a state-of-the-art 2D hair manipulation network. MOST-GAN achieves photorealistic manipulation of portrait images with fully disentangled 3D control over their physical attributes, enabling extreme manipulation of lighting, facial expression, and pose variations up to full profile view.

Liang Chen,1 Xiaoming Liu,2 Hongjie Tang1 1School of Journalism and Communication, Tsinghua University, Beijing, People’s Republic of China; 2Department of Journalism, School of Communication, Hong Kong Baptist University, Hong Kong, People’s Republic of ChinaCorrespondence: Xiaoming Liu, Department of Journalism, School of Communication, Hong Kong Baptist University, 5 Hereford Road, Kowloon Tong, Hong Kong, People’s Republic of China, Tel +852 59826538, Email liuxiaoming@hkbu.edu.hkPurpose: Applying the parental mediation theory, this study investigates the main effects and interactive effects of three parental social media strategies, ie, active mediation, restrictive mediation, and non-intrusive inspection, on cyberbullying among teenagers. A matched child-parent survey was conducted with 642 secondary-school students aged 13– 18 and their parents in China.Results: The results showed that active mediation was negatively associated with both cyberbullying victimization and perpetration. Restrictive mediation was not significantly associated with cyberbullying. Non-intrusive inspection was positively associated with cyberbullying perpetration but not associated with cyberbullying victimization. Moreover, the two-way interaction between active and restrictive mediation as well as the three-way interaction between the three parental mediation strategies significantly affected cyberbullying perpetration and victimization. Specifically, the combination of high-level active mediation and non-intrusive inspection with low-level restrictive mediation was the most effective strategy for preventing cyberbullying.Conclusion: This finding significantly contributes to the parental mediation literature and provide theoretical guidelines for parental intervention to prevent cyberbullying among teenager.Keywords: parental mediation, social media, cyberbullying, interactive effects

The complex and rigid floc structure often limits the reutilization of waste activated sludge (WAS). Electrochemical pretreatment (EPT) is one of the most effective technologies that can enhance WAS disintegration. But a comprehensive investigation into how multiple EPT conditions work was rarely reported. The study evaluated the effects of multiple EPT conditions, i.e., different electrolytes (NaCl, Na2SO4, and CaCl2), electrolytes dosage (0 g/L, 0.5 g/L, 1.0 g/L, and 3.0 g/L), EPT current (0 A, 0.5 A, 1.0 A, and 3.0 A) and EPT time (0 min, 30 min, 60 min, and 90 min) on WAS disintegration. The results showed that NaCl was outstanding from other electrolytes in promoting more WAS disintegration. Besides, a relatively higher NaCl dosage, a higher EPT current, and a longer EPT time promoted more reactive chlorine species (RCS), thus enhancing WAS disintegration in terms of extracellular polymeric substances (EPS) structure destruction and biodegradable organic matter release. After EPT for 60 min at NaCl dosage of 1.0 g/L and current of 1.0 A, the EPS multilayer structure destruction, biodegradable organic matters release, and soluble chemical oxygen demand (SCOD) increase in the supernatant were enhanced by 17.2 %, 130.5 %, and 238.7 %, respectively. Then a predictive quadratic model was established and the impact significance of the above EPT factors for enhancing WAS disintegration followed dosage of NaCl > current > EPT time. Furthermore, response surface methodology (RSM) suggested NaCl dosage of 2.75 g/L, current of 2.0 A, and EPT time of 30 min were the optimal EPT conditions, bringing a 42.0 % increase in the net economic benefit of WAS treatment compared to without EPT. © 2023 Elsevier B.V.

Generating synthetic datasets for training face recognition models is challenging because dataset generation entails more than creating high fidelity images. It involves generating multiple images of same subjects under different factors (\textit{e.g.}, variations in pose, illumination, expression, aging and occlusion) which follows the real image conditional distribution. Previous works have studied the generation of synthetic datasets using GAN or 3D models. In this work, we approach the problem from the aspect of combining subject appearance (ID) and external factor (style) conditions. These two conditions provide a direct way to control the inter-class and intra-class variations. To this end, we propose a Dual Condition Face Generator (DCFace) based on a diffusion model. Our novel Patch-wise style extractor and Time-step dependent ID loss enables DCFace to consistently produce face images of the same subject under different styles with precise control. Face recognition models trained on synthetic images from the proposed DCFace provide higher verification accuracies compared to previous works by $6.11\%$ on average in $4$ out of $5$ test datasets, LFW, CFP-FP, CPLFW, AgeDB and CALFW. Code is available at https://github.com/mk-minchul/dcface, Comment: To appear in CVPR 2023

Dense geometric matching determines the dense pixel-wise correspondence between a source and support image corresponding to the same 3D structure. Prior works employ an encoder of transformer blocks to correlate the two-frame features. However, existing monocular pretraining tasks, e.g., image classification, and masked image modeling (MIM), can not pretrain the cross-frame module, yielding less optimal performance. To resolve this, we reformulate the MIM from reconstructing a single masked image to reconstructing a pair of masked images, enabling the pretraining of transformer module. Additionally, we incorporate a decoder into pretraining for improved upsampling results. Further, to be robust to the textureless area, we propose a novel cross-frame global matching module (CFGM). Since the most textureless area is planar surfaces, we propose a homography loss to further regularize its learning. Combined together, we achieve the State-of-The-Art (SoTA) performance on geometric matching. Codes and models are available at https://github.com/ShngJZ/PMatch., Comment: CVPR2023

Advancements in the generation quality of various Generative Models (GMs) has made it necessary to not only perform binary manipulation detection but also localize the modified pixels in an image. However, prior works termed as passive for manipulation localization exhibit poor generalization performance over unseen GMs and attribute modifications. To combat this issue, we propose a proactive scheme for manipulation localization, termed MaLP. We encrypt the real images by adding a learned template. If the image is manipulated by any GM, this added protection from the template not only aids binary detection but also helps in identifying the pixels modified by the GM. The template is learned by leveraging local and global-level features estimated by a two-branch architecture. We show that MaLP performs better than prior passive works. We also show the generalizability of MaLP by testing on 22 different GMs, providing a benchmark for future research on manipulation localization. Finally, we show that MaLP can be used as a discriminator for improving the generation quality of GMs. Our models/codes are available at www.github.com/vishal3477/pro_loc., Comment: Published at Conference on Computer Vision and Pattern Recognition 2023

Differences in forgery attributes of images generated in CNN-synthesized and image-editing domains are large, and such differences make a unified image forgery detection and localization (IFDL) challenging. To this end, we present a hierarchical fine-grained formulation for IFDL representation learning. Specifically, we first represent forgery attributes of a manipulated image with multiple labels at different levels. Then we perform fine-grained classification at these levels using the hierarchical dependency between them. As a result, the algorithm is encouraged to learn both comprehensive features and inherent hierarchical nature of different forgery attributes, thereby improving the IFDL representation. Our proposed IFDL framework contains three components: multi-branch feature extractor, localization and classification modules. Each branch of the feature extractor learns to classify forgery attributes at one level, while localization and classification modules segment the pixel-level forgery region and detect image-level forgery, respectively. Lastly, we construct a hierarchical fine-grained dataset to facilitate our study. We demonstrate the effectiveness of our method on $7$ different benchmarks, for both tasks of IFDL and forgery attribute classification. Our source code and dataset can be found: \href{https://github.com/CHELSEA234/HiFi_IFDL}{github.com/CHELSEA234/HiFi-IFDL}., Comment: To appear at CVPR2023; 17 pages, 15 figures and 10 tables

This work studies the generalization issue of face anti-spoofing (FAS) models on domain gaps, such as image resolution, blurriness and sensor variations. Most prior works regard domain-specific signals as a negative impact, and apply metric learning or adversarial losses to remove them from feature representation. Though learning a domain-invariant feature space is viable for the training data, we show that the feature shift still exists in an unseen test domain, which backfires on the generalizability of the classifier. In this work, instead of constructing a domain-invariant feature space, we encourage domain separability while aligning the live-to-spoof transition (i.e., the trajectory from live to spoof) to be the same for all domains. We formulate this FAS strategy of separability and alignment (SA-FAS) as a problem of invariant risk minimization (IRM), and learn domain-variant feature representation but domain-invariant classifier. We demonstrate the effectiveness of SA-FAS on challenging cross-domain FAS datasets and establish state-of-the-art performance., Comment: Accepted in CVPR2023

N6-methyladenosine (m6A) is the most abundant internal modification on RNA. It is a dynamical and reversible process, which is regulated by m6A methyltransferase and m6A demethylase. The m6A modified RNA can be specifically recognized by the m6A reader, leading to RNA splicing, maturation, degradation or translation. The abnormality of m6A RNA modification is closely related to a variety of biological processes, especially the occurrence and development of tumors. Recent studies have shown that m6A RNA modification is involved in the pathogenesis of skin cancers. However, the precise molecular mechanisms of m6A-mediated cutaneous tumorigenesis have not been fully elucidated. Therefore, this review will summarize the biological characteristics of m6A modification, its regulatory role and mechanism in skin cancers, and the recent research progress of m6A-related molecular drugs, aiming to provide new ideas for clinical diagnosis and targeted therapy of cutaneous cancers.

Effects of 0–90% decalcification on the coagulation behavior and gastrointestinal digestibility of bovine casein micelles were studied in vitro under infant, adult and elderly conditions. Gastric coagula became looser with increasing level of decalcification, especially up to 40% under the infant condition and up to 69% under the adult and elderly conditions. Casein degradation showed a similarly increasing trend with increasing level of decalcification under the three age conditions. Casein degradation was fastest under the adult condition, but slowest under the elderly and infant conditions at 0–40 and 69–90% decalcification, respectively. Formation of free amino groups and small peptides increased with increasing level of decalcification and in order of infant, elderly and adult conditions. In intestinal phase, the remaining caseins were very rapidly degraded. Formation of free amino groups and smaller peptides showed minor differences with increasing level of decalcification, but increased in order of infant, elderly and adult conditions. This study highlighted the different level of decalcification suitable for diverse age populations to improve the gastrointestinal digestibility of bovine casein micelles.

This work presents FaceX framework, a novel facial generalist model capable of handling diverse facial tasks simultaneously. To achieve this goal, we initially formulate a unified facial representation for a broad spectrum of facial editing tasks, which macroscopically decomposes a face into fundamental identity, intra-personal variation, and environmental factors. Based on this, we introduce Facial Omni-Representation Decomposing (FORD) for seamless manipulation of various facial components, microscopically decomposing the core aspects of most facial editing tasks. Furthermore, by leveraging the prior of a pretrained StableDiffusion (SD) to enhance generation quality and accelerate training, we design Facial Omni-Representation Steering (FORS) to first assemble unified facial representations and then effectively steer the SD-aware generation process by the efficient Facial Representation Controller (FRC). %Without any additional features, Our versatile FaceX achieves competitive performance compared to elaborate task-specific models on popular facial editing tasks. Full codes and models will be available at https://github.com/diffusion-facex/FaceX., Comment: Project page: https://diffusion-facex.github.io

We propose INFAMOUS-NeRF, an implicit morphable face model that introduces hypernetworks to NeRF to improve the representation power in the presence of many training subjects. At the same time, INFAMOUS-NeRF resolves the classic hypernetwork tradeoff of representation power and editability by learning semantically-aligned latent spaces despite the subject-specific models, all without requiring a large pretrained model. INFAMOUS-NeRF further introduces a novel constraint to improve NeRF rendering along the face boundary. Our constraint can leverage photometric surface rendering and multi-view supervision to guide surface color prediction and improve rendering near the surface. Finally, we introduce a novel, loss-guided adaptive sampling method for more effective NeRF training by reducing the sampling redundancy. We show quantitatively and qualitatively that our method achieves higher representation power than prior face modeling methods in both controlled and in-the-wild settings. Code and models will be released upon publication.

Model Parsing defines the research task of predicting hyperparameters of the generative model (GM), given a generated image as input. Since a diverse set of hyperparameters is jointly employed by the generative model, and dependencies often exist among them, it is crucial to learn these hyperparameter dependencies for the improved model parsing performance. To explore such important dependencies, we propose a novel model parsing method called Learnable Graph Pooling Network (LGPN). Specifically, we transform model parsing into a graph node classification task, using graph nodes and edges to represent hyperparameters and their dependencies, respectively. Furthermore, LGPN incorporates a learnable pooling-unpooling mechanism tailored to model parsing, which adaptively learns hyperparameter dependencies of GMs used to generate the input image. We also extend our proposed method to CNN-generated image detection and coordinate attacks detection. Empirically, we achieve state-of-the-art results in model parsing and its extended applications, showing the effectiveness of our method. Our source code are available., Comment: 24 pages, 15 figures, 17 tables

Despite the widespread adoption of face recognition technology around the world, and its remarkable performance on current benchmarks, there are still several challenges that must be covered in more detail. This paper offers an overview of the Face Recognition Challenge in the Era of Synthetic Data (FRCSyn) organized at WACV 2024. This is the first international challenge aiming to explore the use of synthetic data in face recognition to address existing limitations in the technology. Specifically, the FRCSyn Challenge targets concerns related to data privacy issues, demographic biases, generalization to unseen scenarios, and performance limitations in challenging scenarios, including significant age disparities between enrollment and testing, pose variations, and occlusions. The results achieved in the FRCSyn Challenge, together with the proposed benchmark, contribute significantly to the application of synthetic data to improve face recognition technology., Comment: 10 pages, 1 figure, WACV 2024 Workshops

Existing LLM-based systems for writing long-form stories or story outlines frequently suffer from unnatural pacing, whether glossing over important events or over-elaborating on insignificant details, resulting in a jarring experience for the reader. We propose a CONCrete Outline ConTrol (CONCOCT) system to improve pacing when automatically generating story outlines. We first train a concreteness evaluator to judge which of two events is more concrete (low-level-detailed). This evaluator can then be used to control pacing in hierarchical outline generation; in this work, we explore a vaguest-first expansion procedure that aims for uniform pacing. We further use the evaluator to filter new outline items based on predicted concreteness. Compared to a baseline hierarchical outline generator, humans judge CONCOCT's pacing to be more consistent over 57% of the time across multiple outline lengths; the gains also translate to downstream stories. All code, data, and models are open-sourced., Comment: EMNLP Findings 2023

This paper presents a novel approach, called Prototype-based Self-Distillation (ProS), for unsupervised face representation learning. The existing supervised methods heavily rely on a large amount of annotated training facial data, which poses challenges in terms of data collection and privacy concerns. To address these issues, we propose ProS, which leverages a vast collection of unlabeled face images to learn a comprehensive facial omni-representation. In particular, ProS consists of two vision-transformers (teacher and student models) that are trained with different augmented images (cropping, blurring, coloring, etc.). Besides, we build a face-aware retrieval system along with augmentations to obtain the curated images comprising predominantly facial areas. To enhance the discrimination of learned features, we introduce a prototype-based matching loss that aligns the similarity distributions between features (teacher or student) and a set of learnable prototypes. After pre-training, the teacher vision transformer serves as a backbone for downstream tasks, including attribute estimation, expression recognition, and landmark alignment, achieved through simple fine-tuning with additional layers. Extensive experiments demonstrate that our method achieves state-of-the-art performance on various tasks, both in full and few-shot settings. Furthermore, we investigate pre-training with synthetic face images, and ProS exhibits promising performance in this scenario as well., Comment: This paper has been accepted in WACV2024

The zero-shot open-vocabulary challenge in image classification is tackled by pretrained vision-language models like CLIP, which benefit from incorporating class-specific knowledge from large language models (LLMs) like ChatGPT. However, biases in CLIP lead to similar descriptions for distinct but related classes, prompting our novel image classification framework via hierarchical comparisons: using LLMs to recursively group classes into hierarchies and classifying images by comparing image-text embeddings at each hierarchy level, resulting in an intuitive, effective, and explainable approach., Comment: Neurips 2023 Poster

Previous research in $2D$ object detection focuses on various tasks, including detecting objects in generic and camouflaged images. These works are regarded as passive works for object detection as they take the input image as is. However, convergence to global minima is not guaranteed to be optimal in neural networks; therefore, we argue that the trained weights in the object detector are not optimal. To rectify this problem, we propose a wrapper based on proactive schemes, PrObeD, which enhances the performance of these object detectors by learning a signal. PrObeD consists of an encoder-decoder architecture, where the encoder network generates an image-dependent signal termed templates to encrypt the input images, and the decoder recovers this template from the encrypted images. We propose that learning the optimum template results in an object detector with an improved detection performance. The template acts as a mask to the input images to highlight semantics useful for the object detector. Finetuning the object detector with these encrypted images enhances the detection performance for both generic and camouflaged. Our experiments on MS-COCO, CAMO, COD$10$K, and NC$4$K datasets show improvement over different detectors after applying PrObeD. Our models/codes are available at https://github.com/vishal3477/Proactive-Object-Detection., Comment: Accepted at Neurips 2023

Long-Term Person Re-Identification (LT-ReID) has become increasingly crucial in computer vision and biometrics. In this work, we aim to extend LT-ReID beyond pedestrian recognition to include a wider range of real-world human activities while still accounting for cloth-changing scenarios over large time gaps. This setting poses additional challenges due to the geometric misalignment and appearance ambiguity caused by the diversity of human pose and clothing. To address these challenges, we propose a new approach 3DInvarReID for (i) disentangling identity from non-identity components (pose, clothing shape, and texture) of 3D clothed humans, and (ii) reconstructing accurate 3D clothed body shapes and learning discriminative features of naked body shapes for person ReID in a joint manner. To better evaluate our study of LT-ReID, we collect a real-world dataset called CCDA, which contains a wide variety of human activities and clothing changes. Experimentally, we show the superior performance of our approach for person ReID., Comment: 10 pages, 7 figures, accepted by ICCV 2023

Prompt-based pre-trained language models (PLMs) paradigm have succeeded substantially in few-shot natural language processing (NLP) tasks. However, prior discrete prompt optimization methods require expert knowledge to design the base prompt set and identify high-quality prompts, which is costly, inefficient, and subjective. Meanwhile, existing continuous prompt optimization methods improve the performance by learning the ideal prompts through the gradient information of PLMs, whose high computational cost, and low readability and generalizability are often concerning. To address the research gap, we propose a Dialogue-comprised Policy-gradient-based Discrete Prompt Optimization ($DP_2O$) method. We first design a multi-round dialogue alignment strategy for readability prompt set generation based on GPT-4. Furthermore, we propose an efficient prompt screening metric to identify high-quality prompts with linear complexity. Finally, we construct a reinforcement learning (RL) framework based on policy gradients to match the prompts to inputs optimally. By training a policy network with only 0.67% of the PLM parameter size on the tasks in the few-shot setting, $DP_2O$ outperforms the state-of-the-art (SOTA) method by 1.52% in accuracy on average on four open-source datasets. Moreover, subsequent experiments also demonstrate that $DP_2O$ has good universality, robustness, and generalization ability., Comment: AAAI 2024 Main Track

Whole-body biometric recognition is an important area of research due to its vast applications in law enforcement, border security, and surveillance. This paper presents the end-to-end design, development and evaluation of FarSight, an innovative software system designed for whole-body (fusion of face, gait and body shape) biometric recognition. FarSight accepts videos from elevated platforms and drones as input and outputs a candidate list of identities from a gallery. The system is designed to address several challenges, including (i) low-quality imagery, (ii) large yaw and pitch angles, (iii) robust feature extraction to accommodate large intra-person variabilities and large inter-person similarities, and (iv) the large domain gap between training and test sets. FarSight combines the physics of imaging and deep learning models to enhance image restoration and biometric feature encoding. We test FarSight's effectiveness using the newly acquired IARPA Biometric Recognition and Identification at Altitude and Range (BRIAR) dataset. Notably, FarSight demonstrated a substantial performance increase on the BRIAR dataset, with gains of +11.82% Rank-20 identification and +11.3% TAR@1% FAR., Comment: 11 pages, 7 figures, accepted in WACV 2024

3D sensing for monocular in-the-wild images, e.g., depth estimation and 3D object detection, has become increasingly important. However, the unknown intrinsic parameter hinders their development and deployment. Previous methods for the monocular camera calibration rely on specific 3D objects or strong geometry prior, such as using a checkerboard or imposing a Manhattan World assumption. This work solves the problem from the other perspective by exploiting the monocular 3D prior. Our method is assumption-free and calibrates the complete $4$ Degree-of-Freedom (DoF) intrinsic parameters. First, we demonstrate intrinsic is solved from two well-studied monocular priors, i.e., monocular depthmap, and surface normal map. However, this solution imposes a low-bias and low-variance requirement for depth estimation. Alternatively, we introduce a novel monocular 3D prior, the incidence field, defined as the incidence rays between points in 3D space and pixels in the 2D imaging plane. The incidence field is a pixel-wise parametrization of the intrinsic invariant to image cropping and resizing. With the estimated incidence field, a robust RANSAC algorithm recovers intrinsic. We demonstrate the effectiveness of our method by showing superior performance on synthetic and zero-shot testing datasets. Beyond calibration, we demonstrate downstream applications in image manipulation detection & restoration, uncalibrated two-view pose estimation, and 3D sensing. Codes, models, and data will be held in https://github.com/ShngJZ/WildCamera.

Crystalline carbon nitride is regarded as the new generation of emerging metal-free photocatalysts as opposed to polymeric carbon nitride (g-CN) because of its high crystalline structure and ultrahigh photocatalytic water splitting performance. However, further advances in crystalline g-CN are significantly restricted by the sluggish separation of charge carriers and limited active sites. In this study, we demonstrate the successful synthesis of heptazine-triazine donor-acceptor-based ultrathin crystalline g-CN nanosheets (UCCN) using a combined hot air exfoliation and molten salt (NaCl/KCl) copolymerization approach. The synergy of the donor-acceptor heterojunction and the ultrathin structure greatly accelerated the separation of the charge carriers and enriched the active sites. Accordingly, the superior hydrogen evolution activity and an ultrahigh apparent quantum efficiency of 73.6% at 420 nm under a natural photosynthetic environment were achieved by UCCN, positioning this material at the top among reported conjugated g-CN materials. This study provides a novel paradigm for the development of donor-acceptor-based ultrathin crystalline layered materials.

N6-methyladenosine (m6A) is the most abundant internal modification on RNA. It is a dynamical and reversible process, which is regulated by m6A methyltransferase and m6A demethylase. The m6A modified RNA can be specifically recognized by the m6A reader, leading to RNA splicing, maturation, degradation or translation. The abnormality of m6A RNA modification is closely related to a variety of biological processes, especially the occurrence and development of tumors. Recent studies have shown that m6A RNA modification is involved in the pathogenesis of skin cancers. However, the precise molecular mechanisms of m6A-mediated cutaneous tumorigenesis have not been fully elucidated. Therefore, this review will summarize the biological characteristics of m6A modification, its regulatory role and mechanism in skin cancers, and the recent research progress of m6A-related molecular drugs, aiming to provide new ideas for clinical diagnosis and targeted therapy of cutaneous cancers.

The human lung differs substantially from its mouse counterpart, resulting in a distinct distal airway architecture affected by disease pathology in chronic obstructive pulmonary disease. In humans, the distal branches of the airway interweave with the alveolar gas-exchange niche, forming an anatomical structure known as the respiratory bronchioles. Owing to the lack of a counterpart in mouse, the cellular and molecular mechanisms that govern respiratory bronchioles in the human lung remain uncharacterized. Here we show that human respiratory bronchioles contain a unique secretory cell population that is distinct from cells in larger proximal airways. Organoid modelling reveals that these respiratory airway secretory (RAS) cells act as unidirectional progenitors for alveolar type 2 cells, which are essential for maintaining and regenerating the alveolar niche. RAS cell lineage differentiation into alveolar type 2 cells is regulated by Notch and Wnt signalling. In chronic obstructive pulmonary disease, RAS cells are altered transcriptionally, corresponding to abnormal alveolar type 2 cell states, which are associated with smoking exposure in both humans and ferrets. These data identify a distinct progenitor in a region of the human lung that is not found in mouse that has a critical role in maintaining the gas-exchange compartment and is altered in chronic lung disease.

The objective of this paper is to learn dense 3D shape correspondence for topology-varying generic objects in an unsupervised manner. Conventional implicit functions estimate the occupancy of a 3D point given a shape latent code. Instead, our novel implicit function produces a probabilistic embedding to represent each 3D point in a part embedding space. Assuming the corresponding points are similar in the embedding space, we implement dense correspondence through an inverse function mapping from the part embedding vector to a corresponded 3D point. Both functions are jointly learned with several effective and uncertainty-aware loss functions to realize our assumption, together with the encoder generating the shape latent code. During inference, if a user selects an arbitrary point on the source shape, our algorithm can automatically generate a confidence score indicating whether there is a correspondence on the target shape, as well as the corresponding semantic point if there is one. Such a mechanism inherently benefits man-made objects with different part constitutions. The effectiveness of our approach is demonstrated through unsupervised 3D semantic correspondence and shape segmentation., Comment: arXiv admin note: substantial text overlap with arXiv:2010.12320

Machine-Generated Text (MGT) detection, a task that discriminates MGT from Human-Written Text (HWT), plays a crucial role in preventing misuse of text generative models, which excel in mimicking human writing style recently. Latest proposed detectors usually take coarse text sequences as input and fine-tune pretrained models with standard cross-entropy loss. However, these methods fail to consider the linguistic structure of texts. Moreover, they lack the ability to handle the low-resource problem which could often happen in practice considering the enormous amount of textual data online. In this paper, we present a coherence-based contrastive learning model named CoCo to detect the possible MGT under low-resource scenario. To exploit the linguistic feature, we encode coherence information in form of graph into text representation. To tackle the challenges of low data resource, we employ a contrastive learning framework and propose an improved contrastive loss for preventing performance degradation brought by simple samples. The experiment results on two public datasets and two self-constructed datasets prove our approach outperforms the state-of-art methods significantly. Also, we surprisingly find that MGTs originated from up-to-date language models could be easier to detect than these from previous models, in our experiments. And we propose some preliminary explanations for this counter-intuitive phenomena. All the codes and datasets are open-sourced., Comment: Accepted by EMNLP 2023 main cofference

Feature fusion plays a crucial role in unconstrained face recognition where inputs (probes) comprise of a set of $N$ low quality images whose individual qualities vary. Advances in attention and recurrent modules have led to feature fusion that can model the relationship among the images in the input set. However, attention mechanisms cannot scale to large $N$ due to their quadratic complexity and recurrent modules suffer from input order sensitivity. We propose a two-stage feature fusion paradigm, Cluster and Aggregate, that can both scale to large $N$ and maintain the ability to perform sequential inference with order invariance. Specifically, Cluster stage is a linear assignment of $N$ inputs to $M$ global cluster centers, and Aggregation stage is a fusion over $M$ clustered features. The clustered features play an integral role when the inputs are sequential as they can serve as a summarization of past features. By leveraging the order-invariance of incremental averaging operation, we design an update rule that achieves batch-order invariance, which guarantees that the contributions of early image in the sequence do not diminish as time steps increase. Experiments on IJB-B and IJB-S benchmark datasets show the superiority of the proposed two-stage paradigm in unconstrained face recognition. Code and pretrained models are available in https://github.com/mk-minchul/caface, Comment: To appear in NeurIPS 2022

In this work, we study multi-domain learning for face anti-spoofing(MD-FAS), where a pre-trained FAS model needs to be updated to perform equally well on both source and target domains while only using target domain data for updating. We present a new model for MD-FAS, which addresses the forgetting issue when learning new domain data, while possessing a high level of adaptability. First, we devise a simple yet effective module, called spoof region estimator(SRE), to identify spoof traces in the spoof image. Such spoof traces reflect the source pre-trained model's responses that help upgraded models combat catastrophic forgetting during updating. Unlike prior works that estimate spoof traces which generate multiple outputs or a low-resolution binary mask, SRE produces one single, detailed pixel-wise estimate in an unsupervised manner. Secondly, we propose a novel framework, named FAS-wrapper, which transfers knowledge from the pre-trained models and seamlessly integrates with different FAS models. Lastly, to help the community further advance MD-FAS, we construct a new benchmark based on SIW, SIW-Mv2 and Oulu-NPU, and introduce four distinct protocols for evaluation, where source and target domains are different in terms of spoof type, age, ethnicity, and illumination. Our proposed method achieves superior performance on the MD-FAS benchmark than previous methods. Our code and newly curated SIW-Mv2 are publicly available., Comment: To appear at ECCV 2022 as an oral presentation. The SiW-Mv2 dataset is detailed in the Appendix

Although significant advances have been made in face recognition (FR), FR in unconstrained environments remains challenging due to the domain gap between the semi-constrained training datasets and unconstrained testing scenarios. To address this problem, we propose a controllable face synthesis model (CFSM) that can mimic the distribution of target datasets in a style latent space. CFSM learns a linear subspace with orthogonal bases in the style latent space with precise control over the diversity and degree of synthesis. Furthermore, the pre-trained synthesis model can be guided by the FR model, making the resulting images more beneficial for FR model training. Besides, target dataset distributions are characterized by the learned orthogonal bases, which can be utilized to measure the distributional similarity among face datasets. Our approach yields significant performance gains on unconstrained benchmarks, such as IJB-B, IJB-C, TinyFace and IJB-S (+5.76% Rank1)., Comment: to be published in ECCV 2022