24 results on '"Yang, Zongxin"'
Search Results
2. Reshaping the tumor immune microenvironment to improve CAR-T cell-based cancer immunotherapy
- Author
-
Xia, Xueting, Yang, Zongxin, Lu, Qisi, Liu, Zhenyun, Wang, Lei, Du, Jinwen, Li, Yuhua, Yang, Dong-Hua, and Wu, Shaojie
- Published
- 2024
- Full Text
- View/download PDF
3. Integration of the transcriptome and proteome provides insights into the mechanism calcium regulated of Ulva prolifera in response to high-temperature stress
- Author
-
Fan, Meihua, Tang, Xiaowen, Yang, Zongxin, Wang, Jianxin, Zhang, Xiaolin, Yan, Xiaojun, Li, Peng, Xu, Nianjun, and Liao, Zhi
- Published
- 2022
- Full Text
- View/download PDF
4. The Schwarzian Derivative of Harmonic Mappings in the Plane
- Author
-
Nie, Liping and Yang, Zongxin
- Published
- 2020
- Full Text
- View/download PDF
5. TransHuman: A Transformer-based Human Representation for Generalizable Neural Human Rendering
- Author
-
Pan, Xiao, Yang, Zongxin, Ma, Jianxin, Zhou, Chang, and Yang, Yi
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
In this paper, we focus on the task of generalizable neural human rendering which trains conditional Neural Radiance Fields (NeRF) from multi-view videos of different characters. To handle the dynamic human motion, previous methods have primarily used a SparseConvNet (SPC)-based human representation to process the painted SMPL. However, such SPC-based representation i) optimizes under the volatile observation space which leads to the pose-misalignment between training and inference stages, and ii) lacks the global relationships among human parts that is critical for handling the incomplete painted SMPL. Tackling these issues, we present a brand-new framework named TransHuman, which learns the painted SMPL under the canonical space and captures the global relationships between human parts with transformers. Specifically, TransHuman is mainly composed of Transformer-based Human Encoding (TransHE), Deformable Partial Radiance Fields (DPaRF), and Fine-grained Detail Integration (FDI). TransHE first processes the painted SMPL under the canonical space via transformers for capturing the global relationships between human parts. Then, DPaRF binds each output token with a deformable radiance field for encoding the query point under the observation space. Finally, the FDI is employed to further integrate fine-grained information from reference images. Extensive experiments on ZJU-MoCap and H36M show that our TransHuman achieves a significantly new state-of-the-art performance with high efficiency. Project page: https://pansanity666.github.io/TransHuman/, Comment: Accepted by ICCV 2023
- Published
- 2023
6. Video Object Segmentation in Panoptic Wild Scenes
- Author
-
Xu, Yuanyou, Yang, Zongxin, and Yang, Yi
- Subjects
FOS: Computer and information sciences ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition - Abstract
In this paper, we introduce semi-supervised video object segmentation (VOS) to panoptic wild scenes and present a large-scale benchmark as well as a baseline method for it. Previous benchmarks for VOS with sparse annotations are not sufficient to train or evaluate a model that needs to process all possible objects in real-world scenarios. Our new benchmark (VIPOSeg) contains exhaustive object annotations and covers various real-world object categories which are carefully divided into subsets of thing/stuff and seen/unseen classes for comprehensive evaluation. Considering the challenges in panoptic VOS, we propose a strong baseline method named panoptic object association with transformers (PAOT), which uses panoptic identification to associate objects with a pyramid architecture on multiple scales. Experimental results show that VIPOSeg can not only boost the performance of VOS models by panoptic training but also evaluate them comprehensively in panoptic scenes. Previous methods for classic VOS still need to improve in performance and efficiency when dealing with panoptic scenes, while our PAOT achieves SOTA performance with good efficiency on VIPOSeg and previous VOS benchmarks. PAOT also ranks 1st in the VOT2022 challenge. Our dataset is available at https://github.com/yoxu515/VIPOSeg-Benchmark., Accepted to IJCAI2023
- Published
- 2023
7. Global-to-Local Modeling for Video-based 3D Human Pose and Shape Estimation
- Author
-
Shen, Xiaolong, Yang, Zongxin, Wang, Xiaohan, Ma, Jianxin, Zhou, Chang, and Yang, Yi
- Subjects
FOS: Computer and information sciences ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Video-based 3D human pose and shape estimations are evaluated by intra-frame accuracy and inter-frame smoothness. Although these two metrics are responsible for different ranges of temporal consistency, existing state-of-the-art methods treat them as a unified problem and use monotonous modeling structures (e.g., RNN or attention-based block) to design their networks. However, using a single kind of modeling structure is difficult to balance the learning of short-term and long-term temporal correlations, and may bias the network to one of them, leading to undesirable predictions like global location shift, temporal inconsistency, and insufficient local details. To solve these problems, we propose to structurally decouple the modeling of long-term and short-term correlations in an end-to-end framework, Global-to-Local Transformer (GLoT). First, a global transformer is introduced with a Masked Pose and Shape Estimation strategy for long-term modeling. The strategy stimulates the global transformer to learn more inter-frame correlations by randomly masking the features of several frames. Second, a local transformer is responsible for exploiting local details on the human mesh and interacting with the global transformer by leveraging cross-attention. Moreover, a Hierarchical Spatial Correlation Regressor is further introduced to refine intra-frame estimations by decoupled global-local representation and implicit kinematic constraints. Our GLoT surpasses previous state-of-the-art methods with the lowest model parameters on popular benchmarks, i.e., 3DPW, MPI-INF-3DHP, and Human3.6M. Codes are available at https://github.com/sxl142/GLoT.
- Published
- 2023
8. Decoupling Features in Hierarchical Propagation for Video Object Segmentation
- Author
-
Yang, Zongxin and Yang, Yi
- Subjects
FOS: Computer and information sciences ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition - Abstract
This paper focuses on developing a more effective method of hierarchical propagation for semi-supervised Video Object Segmentation (VOS). Based on vision transformers, the recently-developed Associating Objects with Transformers (AOT) approach introduces hierarchical propagation into VOS and has shown promising results. The hierarchical propagation can gradually propagate information from past frames to the current frame and transfer the current frame feature from object-agnostic to object-specific. However, the increase of object-specific information will inevitably lead to the loss of object-agnostic visual information in deep propagation layers. To solve such a problem and further facilitate the learning of visual embeddings, this paper proposes a Decoupling Features in Hierarchical Propagation (DeAOT) approach. Firstly, DeAOT decouples the hierarchical propagation of object-agnostic and object-specific embeddings by handling them in two independent branches. Secondly, to compensate for the additional computation from dual-branch propagation, we propose an efficient module for constructing hierarchical propagation, i.e., Gated Propagation Module, which is carefully designed with single-head attention. Extensive experiments show that DeAOT significantly outperforms AOT in both accuracy and efficiency. On YouTube-VOS, DeAOT can achieve 86.0% at 22.4fps and 82.0% at 53.4fps. Without test-time augmentations, we achieve new state-of-the-art performance on four benchmarks, i.e., YouTube-VOS (86.2%), DAVIS 2017 (86.2%), DAVIS 2016 (92.9%), and VOT 2020 (0.622). Project page: https://github.com/z-x-yang/AOT., Accepted by NeurIPS 2022 (Spotlight)
- Published
- 2022
9. V$^2$L: Leveraging Vision and Vision-language Models into Large-scale Product Retrieval
- Author
-
Wang, Wenhao, Sun, Yifan, Yang, Zongxin, and Yang, Yi
- Subjects
FOS: Computer and information sciences ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Product retrieval is of great importance in the ecommerce domain. This paper introduces our 1st-place solution in eBay eProduct Visual Search Challenge (FGVC9), which is featured for an ensemble of about 20 models from vision models and vision-language models. While model ensemble is common, we show that combining the vision models and vision-language models brings particular benefits from their complementarity and is a key factor to our superiority. Specifically, for the vision models, we use a two-stage training pipeline which first learns from the coarse labels provided in the training set and then conducts fine-grained self-supervised training, yielding a coarse-to-fine metric learning manner. For the vision-language models, we use the textual description of the training image as the supervision signals for fine-tuning the image-encoder (feature extractor). With these designs, our solution achieves 0.7623 MAR@10, ranking the first place among all the competitors. The code is available at: \href{https://github.com/WangWenhao0716/V2L}{V$^2$L}., CVPR FGVC9 eBay eProduct Visual Search Challenge Rank 1st solution
- Published
- 2022
10. A novel resilience analysis methodology for airport networks system from the perspective of different epidemic prevention and control policy responses.
- Author
-
Guo, Jiuxia, Yang, Zongxin, Zhong, Qingwei, Sun, Xiaoqian, and Wang, Yinhai
- Subjects
- *
AIRPORTS , *AIR traffic , *EPIDEMICS , *COVID-19 pandemic , *NETWORK performance - Abstract
As the COVID-19 pandemic fades, the aviation industry is entering a fast recovery period. To analyze airport networks' post-pandemic resilience during the recovery process, this paper proposes a Comprehensive Resilience Assessment (CRA) model approach using the airport networks of China, Europe, and the U.S.A as case studies. The impact of COVID-19 on the networks is analyzed after populating the models of these networks with real air traffic data. The results suggest that the pandemic has caused damage to all three networks, although the damages to the network structures of Europe and the U.S.A are more severe than the damage in China. The analysis suggests that China, as the airport network with less network performance change, has a more stable level of resilience. The analysis also shows that the different levels of stringency policy in prevention and control measures during the epidemic directly affected the recovery rate of the network. This paper provides new insights into the impact of the pandemic on airport network resilience. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
11. Molecular characterization of peptidoglycan recognition proteins from Mytilus coruscus.
- Author
-
Liao, Zhi, Yang, Zongxin, Wang, Ying, He, Jianyu, He, Zhijiang, Zhang, Xiaolin, Buttino, Isabella, Qi, Pengzhi, Fan, Meihua, Guo, Baoying, Yan, Xiaojun, and He, Menglan
- Subjects
- *
MYTILUS , *IMMUNE recognition , *MORPHOGENESIS - Abstract
Mytilus shows great immune resistance to various bacteria from the living waters, indicating a complex immune recognition mechanism against various microbes. Peptidoglycan recognition proteins (PGRPs) play an important role in the defense against invading microbes via the recognition of the immunogenic substance peptidoglycan (PGN). Therefore, eight PGRPs were identified from the gill transcriptome of Mytilus coruscus. The sequence features, expression pattern in various organs and larval development stages, and microbes induced expression profiles of these Mytilus PGRPs were determined. Our data revealed the constitutive expression of PGRPs in various organs with relative higher expression level in immune-related organs. The expression of PGRPs is developmentally regulated, and most PGRPs are undetectable in larvae stages. The expression level of most PGRPs was significantly increased with in vivo microbial challenges, showing strong response to Gram-positive strain in gill and digestive gland, strong response to Gram-negative strain in hemocytes, and relative weaker response to fungus in the three tested organs. In addition, the function analysis of the representative recombinant expressed PGRP (rMcPGRP-2) confirmed the antimicrobial and agglutination activities, showing the immune-related importance of PGRP in Mytilus. Our work suggests that Mytilus PGRPs can act as pattern recognition receptors to recognize the invading microorganisms and the antimicrobial effectors during the innate immune response of Mytilus. • Eight PGRP genes were identified from the gill transcriptome of M. coruscus. • The expression profiles of the eight Mytilus PGRP genes were determined. • McPGRPs showed different responses to different in vivo microbial challenges. • Recombinant McPGRP-2 showed classical PGRP functions. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
12. Associating Objects with Transformers for Video Object Segmentation
- Author
-
Yang, Zongxin, Wei, Yunchao, and Yang, Yi
- Subjects
FOS: Computer and information sciences ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition - Abstract
This paper investigates how to realize better and more efficient embedding learning to tackle the semi-supervised video object segmentation under challenging multi-object scenarios. The state-of-the-art methods learn to decode features with a single positive object and thus have to match and segment each target separately under multi-object scenarios, consuming multiple times computing resources. To solve the problem, we propose an Associating Objects with Transformers (AOT) approach to match and decode multiple objects uniformly. In detail, AOT employs an identification mechanism to associate multiple targets into the same high-dimensional embedding space. Thus, we can simultaneously process multiple objects' matching and segmentation decoding as efficiently as processing a single object. For sufficiently modeling multi-object association, a Long Short-Term Transformer is designed for constructing hierarchical matching and propagation. We conduct extensive experiments on both multi-object and single-object benchmarks to examine AOT variant networks with different complexities. Particularly, our R50-AOT-L outperforms all the state-of-the-art competitors on three popular benchmarks, i.e., YouTube-VOS (84.1% J&F), DAVIS 2017 (84.9%), and DAVIS 2016 (91.1%), while keeping more than $3\times$ faster multi-object run-time. Meanwhile, our AOT-T can maintain real-time multi-object speed on the above benchmarks. Based on AOT, we ranked 1st in the 3rd Large-scale VOS Challenge., NeurIPS 2021. 20 pages, 9 figures, 5 tables
- Published
- 2021
13. Pre-Inpainting Convolutional Skip Triple Attention Segmentation Network for AGV Lane Detection in Overexposure Environment.
- Author
-
Yang, Zongxin, Yang, Xu, Wu, Long, Hu, Jiemin, Zou, Bo, Zhang, Yong, and Zhang, Jianlong
- Subjects
DEEP learning ,AUTOMATED guided vehicle systems ,IMAGE segmentation - Abstract
Visual navigation is an important guidance method for industrial automated guided vehicles (AGVs). In the actual guidance, the overexposure environment may be encountered by the AGV lane image, which seriously reduces the accuracy of lane detection. Although the image segmentation method based on deep learning is widely used in lane detection, it cannot solve the problem of overexposure of lane images. At the same time, the requirements of segmentation accuracy and inference speed cannot be met simultaneously by existing segmentation networks. Aiming at the problem of incomplete lane segmentation in an overexposure environment, a lane detection method combining image inpainting and image segmentation is proposed. In this method, the overexposed lane image is repaired and reconstructed by the MAE network, and then the image is input into the image segmentation network for lane segmentation. In addition, a convolutional skip triple attention (CSTA) image segmentation network is proposed. CSTA improves the inference speed of the model under the premise of ensuring high segmentation accuracy. Finally, the lane segmentation performance of the proposed method is evaluated in three image segmentation evaluation metrics (IoU, F
1 -score, and PA) and inference time. Experimental results show that the proposed CSTA network has higher segmentation accuracy and faster inference speed. [ABSTRACT FROM AUTHOR]- Published
- 2022
- Full Text
- View/download PDF
14. Collaborative Video Object Segmentation by Multi-Scale Foreground-Background Integration.
- Author
-
Yang, Zongxin, Wei, Yunchao, and Yang, Yi
- Subjects
- *
CONVOLUTIONAL neural networks - Abstract
This paper investigates the principles of embedding learning to tackle the challenging semi-supervised video object segmentation. Unlike previous practices that focus on exploring the embedding learning of foreground object (s), we consider background should be equally treated. Thus, we propose a Collaborative video object segmentation by Foreground-Background Integration (CFBI) approach. CFBI separates the feature embedding into the foreground object region and its corresponding background region, implicitly promoting them to be more contrastive and improving the segmentation results accordingly. Moreover, CFBI performs both pixel-level matching processes and instance-level attention mechanisms between the reference and the predicted sequence, making CFBI robust to various object scales. Based on CFBI, we introduce a multi-scale matching structure and propose an Atrous Matching strategy, resulting in a more robust and efficient framework, CFBI+. We conduct extensive experiments on two popular benchmarks, i.e., DAVIS and YouTube-VOS. Without applying any simulated data for pre-training, our CFBI+ achieves the performance ($\mathcal {J}$ J & $\mathcal {F}$ F ) of 82.9 and 82.8 percent, outperforming all the other state-of-the-art methods. Code: https://github.com/z-x-yang/CFBI. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
15. Sequence Modelling with Deep Learning for Visual Content Generation and Understanding
- Author
-
Yang, Zongxin
- Subjects
ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION - Abstract
University of Technology Sydney. Faculty of Engineering and Information Technology. Although convolutional neural networks have proven to be effective and stable in image feature learning, sequence modelling is still critical for learning spatial and temporal context information. In an image scenario, different semantic structures can be regarded as a sequence arranged along the horizontal (or vertical) direction. Moreover, in a video scenario, temporal sequence modelling is necessary for understanding inter-frame relationships, such as object movement and occlusion. This thesis explores more effective spatial or temporal sequence modelling for image or video scenario understanding. For the former, an encoder-decoder framework is proposed to split an input scenario into a sequence of spatial features and reconstruct the input. By modelling spatial sequence information, the framework can even predict new scenes with very large scales in length while keeping a consistent style regarding the given input. For video understanding, the thesis processes temporal sequences in a recurrent manner (i.e., frame by frame), which is more memory-efficient. In addition, the thesis proposes to implicitly impose the feature embedding of each target and relative background to be contrastive throughout the temporal sequence, promoting the results of downstream tasks accordingly. Besides, a novel transformation module is designed to model channel relationships for improving intra-frame representation ability. To validate proposed approaches and components, extensive experiments are conducted on image outpainting, instance segmentation, object detection, classification, video classification, and video object segmentation.
- Published
- 2021
16. On the extremal sets of extremal quasiconformal mappings
- Author
-
Zhou, Zemin, Chen, Jixiu, and Yang, Zongxin
- Published
- 2003
- Full Text
- View/download PDF
17. Quantitative method for resilience assessment framework of airport network during COVID-19.
- Author
-
Guo, Jiuxia, Li, Yang, Yang, Zongxin, and Zhu, Xinping
- Subjects
COVID-19 pandemic ,COVID-19 ,QUANTITATIVE research ,AIRPORTS ,AIR traffic - Abstract
The resilience and vulnerability of airport networks are significant challenges during the COVID-19 global pandemic. Previous studies considered node failure of networks under natural disasters and extreme weather. Herein, we propose a complex network methodology combined with data-driven to assess the resilience of airport networks toward global-scale disturbance using the Chinese airport network (CAN) and the European airport network (EAN) as a case study. The assessment framework includes vulnerability and resilience analyses from the network- and node-level perspectives. Subsequently, we apply the framework to analyze the airport networks in China and Europe. Specifically, real air traffic data for 232 airports in China and 82 airports in Europe are selected to form the CAN and EAN, respectively. The complex network analysis reveals that the CAN and the EAN are scale-free small-world networks, that are resilient to random attacks. However, the connectivity and vulnerability of the CAN are inferior to those of the EAN. In addition, we select the passenger throughput from the top-50 airports in China and Europe to perform a comparative analysis. By comparing the resilience evaluation of individual airports, we discovered that the factors of resilience assessment of an airport network for global disturbance considers the network metrics and the effect of government policy in actual operations. Additionally, this study also proves that a country's emergency response-ability towards the COVID-19 has a significantly affectes the recovery of its airport network. [ABSTRACT FROM AUTHOR]
- Published
- 2021
- Full Text
- View/download PDF
18. Synthesis and properties of bio-based cationic photopolymerizable polysiloxane oxacyclobutanes based on nopol.
- Author
-
Hu, Yunlang, Yang, Zongxin, Gao, Yanjing, Nie, Jun, and Sun, Fang
- Subjects
- *
ADDITION polymerization , *GLASS transition temperature , *CONTACT angle , *PHOTOPOLYMERIZATION , *MONOMERS , *BENZOXAZINES - Abstract
Nopol as a member of the terpene family is a biorenewable substrate and used widely for the modification of materials. In this work, a series of bio-based cationic photopolymerizable polysiloxane oxacyclobutanes based on nopol with different silicone chain lengths (Nopol-Si n , n = 3, 6, 9) were synthesized. They show good photopolymerization properties and are able to boost the photopolymerization of aliphatic epoxy monomer 3,4-epoxycyclohexylmethyl 3,4-epoxycyclohexanecarboxylate (E4221). The final conversions of the epoxy group and oxetane group of E4221/Nopol-Si n systems are 79.5% and about 69.0%, respectively. Cured films of E4221/Nopol-Si n systems exhibit better hydrophobicity, thermostability and tensile properties compared with that of pure E4221 due to the saturated rigid bridge ring, and flexible and thermostable polysiloxane chain. The water contact angle of E4221/Nopol-Si n systems reaches up to 98.7o. With the increase of polysiloxane chain length, thermostability and tensile properties of E4221/Nopol-Si n systems are boosted, whereas glass transition temperature (T g) is decreased. Moreover, with the increase of Nopol-Si 9 content, thermostability and tensile properties of E4221/Nopol-Si 9 system present a trend of rise first and then decline, whereas T g is continuously reduced. Furthermore, the cured film of E4221/Nopol-Si 3 system has excellent yellowing resistance. [Display omitted] • Nopol-Si n combines features of bio-based monomer nopol, polysiloxane and oxetane. • Nopol-Si n are able to boost cationic polymerization of aliphatic epoxy monomer. • Nopol-Si n improved thermostability and tensile properties of cured films. • Nopol-Si n enhanced hydrophobicity and yellowing resistance of cured films. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
19. Myticofensin, a novel antimicrobial peptide family identified from Mytilus coruscus.
- Author
-
Liu, Lu, He, Menglan, Yang, Zongxin, Wang, Haodong, Zhang, Xiaolin, He, Jianyu, Buttino, Isabella, Qi, Pengzhi, Yan, Xiaojun, and Liao, Zhi
- Subjects
- *
ANTIMICROBIAL peptides , *CATHELICIDINS , *MYTILUS , *PEPTIDE antibiotics , *DEFENSINS , *GENE expression profiling , *AMINO acid sequence , *PERSONAL names - Abstract
In this study, seven transcripts representing a novel antimicrobial peptide (AMP) family with structural features similar to those of arthropod defensins were identified from Mytilus coruscus. These novel defensins from the Mytilus AMP family were named myticofensins. To explore the possible immune-related functions of these myticofensins, we examined their expression profiles in different tissues and larval stages, as well as in three immune-related tissues under the threat of different microbes. Our data revealed that the seven myticofensins had relatively high expression levels in immune-related tissues. Most myticofensins were undetectable, or had low expression levels, in different larval mussel stages. Additionally, in vivo microbial challenges significantly increased the expression levels of myticofensins in M. coruscus hemocytes, gills, and digestive glands, showing different immune response patterns under challenges from different microbes. Our data indicates that different myticofensins may have different immune functions in different tissues. Furthermore, peptide sequences corresponding to the beta-hairpin, alpha-helix, and N-terminal loop of myticofensin were synthesized and the antimicrobial activities of these peptide fragments were tested. Our data confirms the diversity of defensins in Mytilus and reports the complex regulation of these defensins in the mussel immune response to different microbes in immune-related tissues. The immune system of Mytilus has been studied for years as they are a species with strong environmental adaptations. Our data can be regarded as a step forward in the study of the adaptation of Mytilus spp. to an evolving microbial world. • Seven novel antimicrobial peptides named myticofensin were identified from M. coruscus. • The expression profiles of the myticofensin genes were determined. • Myticofensins showed different responses to different microbial challenges in vivo. • The possible active sites were verified for several myticofensins. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
20. SOME RESULTS OF A NEHARI FAMILY.
- Author
-
ZHU, HUACHENG, YANG, ZONGXIN, and CHEN, JIXIU
- Subjects
- *
MATHEMATICAL formulas , *MATHEMATICAL functions , *MATHEMATICAL inequalities , *HOMEOMORPHISMS , *TOPOLOGICAL spaces - Abstract
The authors study the Nehari family Nb and obtain some important properties for this family. [ABSTRACT FROM AUTHOR]
- Published
- 2005
- Full Text
- View/download PDF
21. Scalable Video Object Segmentation With Identification Mechanism.
- Author
-
Yang Z, Miao J, Wei Y, Wang W, Wang X, and Yang Y
- Abstract
This paper delves into the challenges of achieving scalable and effective multi-object modeling for semi-supervised Video Object Segmentation (VOS). Previous VOS methods decode features with a single positive object, limiting the learning of multi-object representation as they must match and segment each target separately under multi-object scenarios. Additionally, earlier techniques catered to specific application objectives and lacked the flexibility to fulfill different speed-accuracy requirements. To address these problems, we present two innovative approaches, Associating Objects with Transformers (AOT) and Associating Objects with Scalable Transformers (AOST). In pursuing effective multi-object modeling, AOT introduces the IDentification (ID) mechanism to allocate each object a unique identity. This approach enables the network to model the associations among all objects simultaneously, thus facilitating the tracking and segmentation of objects in a single network pass. To address the challenge of inflexible deployment, AOST further integrates scalable long short-term transformers that incorporate scalable supervision and layer-wise ID-based attention. This enables online architecture scalability in VOS for the first time and overcomes ID embeddings' representation limitations. Given the absence of a benchmark for VOS involving densely multi-object annotations, we propose a challenging Video Object Segmentation in the Wild (VOSW) benchmark to validate our approaches. We evaluated various AOT and AOST variants using extensive experiments across VOSW and five commonly used VOS benchmarks, including YouTube-VOS 2018 & 2019 Val, DAVIS-2017 Val & Test, and DAVIS-2016. Our approaches surpass the state-of-the-art competitors and display exceptional efficiency and scalability consistently across all six benchmarks. Moreover, we notably achieved the 1
st position in the 3 rd Large-scale Video Object Segmentation Challenge.- Published
- 2024
- Full Text
- View/download PDF
22. Co-Learning Meets Stitch-Up for Noisy Multi-Label Visual Recognition.
- Author
-
Liang C, Yang Z, Zhu L, and Yang Y
- Abstract
In real-world scenarios, collected and annotated data often exhibit the characteristics of multiple classes and long-tailed distribution. Additionally, label noise is inevitable in large-scale annotations and hinders the applications of learning-based models. Although many deep learning based methods have been proposed for handling long-tailed multi-label recognition or label noise respectively, learning with noisy labels in long-tailed multi-label visual data has not been well-studied because of the complexity of long-tailed distribution entangled with multi-label correlation. To tackle such a critical yet thorny problem, this paper focuses on reducing noise based on some inherent properties of multi-label classification and long-tailed learning under noisy cases. In detail, we propose a Stitch-Up augmentation to synthesize a cleaner sample, which directly reduces multi-label noise by stitching up multiple noisy training samples. Equipped with Stitch-Up, a Heterogeneous Co-Learning framework is further designed to leverage the inconsistency between long-tailed and balanced distributions, yielding cleaner labels for more robust representation learning with noisy long-tailed data. To validate our method, we build two challenging benchmarks, named VOC-MLT-Noise and COCO-MLT-Noise, respectively. Extensive experiments are conducted to demonstrate the effectiveness of our proposed method. Compared to a variety of baselines, our method achieves superior results.
- Published
- 2023
- Full Text
- View/download PDF
23. Collaborative Content-Dependent Modeling: A Return to the Roots of Salient Object Detection.
- Author
-
Jiao S, Goel V, Navasardyan S, Yang Z, Khachatryan L, Yang Y, Wei Y, Zhao Y, and Shi H
- Abstract
Salient object detection (SOD) aims to identify the most visually distinctive object(s) from each given image. Most recent progresses focus on either adding elaborative connections among different convolution blocks or introducing boundary-aware supervision to help achieve better segmentation, which is actually moving away from the essence of SOD, i.e., distinctiveness/salience. This paper goes back to the roots of SOD and investigates the principles of how to identify distinctive object(s) in a more effective and efficient way. Intuitively, the salience of one object should largely depend on its global context within the input image. Based on this, we devise a clean yet effective architecture for SOD, named Collaborative Content-Dependent Networks (CCD-Net). In detail, we propose a collaborative content-dependent head whose parameters are conditioned on the input image's global context information. Within the content-dependent head, a hand-crafted multi-scale (HMS) module and a self-induced (SI) module are carefully designed to collaboratively generate content-aware convolution kernels for prediction. Benefited from the content-dependent head, CCD-Net is capable of leveraging global context to detect distinctive object(s) while keeping a simple encoder-decoder design. Extensive experimental results demonstrate that our CCD-Net achieves state-of-the-art results on various benchmarks. Our architecture is simple and intuitive compared to previous solutions, resulting in competitive characteristics with respect to model complexity, operating efficiency, and segmentation accuracy.
- Published
- 2023
- Full Text
- View/download PDF
24. Clinical significance of inflammatory cytokine and chemokine expression in hand, foot and mouth disease.
- Author
-
Shao P, Wu X, Li H, Wu Z, Yang Z, and Yao H
- Subjects
- Child, Preschool, Female, Hand, Foot and Mouth Disease virology, Humans, Infant, Inflammation blood, Inflammation virology, Male, Chemokines blood, Enterovirus A, Human, Hand, Foot and Mouth Disease blood
- Abstract
The present study examined the relationship between cytokine and chemokine expression and the clinical presentation of hand, foot and mouth disease (HFMD), which is currently unclear. The present study involved 28 patients with mild HFMD, 44 patients with severe HFMD and 26 healthy children. Venous blood was tested for cytokine [interleukin (IL)‑4, IL‑12, IL‑18, tumor necrosis factor‑α (TNF‑α), interferon‑γ (IFN‑γ)] and chemokine expression [IL‑8, regulated on activation, normal T cell expressed and secreted (RANTES), monocyte chemoattractant protein‑1 (MCP‑1) and IFN-γ-inducible protein‑10 (IP‑10)]. Stool samples from the patients were tested for enterovirus 71 (EV71) RNA using reverse transcription-polymerase chain reaction. The results indicated that all cytokine/chemokine levels were increased in patients with severe HFMD compared with in patients with mild HFMD or control subjects. In addition, RANTES, MCP‑1, IL‑4, IL‑12 and IL‑18 levels were higher in mild HFMD patients than in the controls. In patients with severe HFMD, all expression levels (with the exception of IL‑8 and IL‑4) were increased in patients with encephalitis plus pulmonary edema compared with those with encephalitis alone. Furthermore, all levels (with the exception of IL‑8) were increased in EV71‑positive patients compared with EV71‑negative patients. In mild HFMD, all levels (with the exception of IL‑8 and IL‑4) were increased in EV71‑positive patients compared with EV71‑negative patients. However, in severe HFMD, only RANTES, IP‑10 and IFN‑γ levels were increased in EV71‑positive patients compared with EV71‑negative patients. In the EV71‑negative group, all levels were increased in severe HFMD compared with mild HFMD. In the EV71‑positive group, all levels (with the exception of IL‑8) were increased in severe HFMD compared with mild HFMD. These results indicated that cytokines and chemokines participate in HFMD pathogenesis, and may be useful to monitor disease progression and predict prognosis.
- Published
- 2017
- Full Text
- View/download PDF
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.