检索结果-内蒙古大学图书馆

FG-UAP: Feature-Gathering Universal Adversarial Perturbation

学校读者我要写书评

暂无评论

FG-UAP: Feature-Gathering Universal Adversarial Perturbation

International Joint Conference on Neural Networks (IJCNN)

作者： Zhixing Ye Xinwen Cheng Xiaolin Huang Department of Automation Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China

Deep Neural Networks (DNNs) are susceptible to elaborately designed perturbations, whether such perturbations are dependent or independent of images. The latter one, called Universal Adversarial Perturbation (UAP), is very attractive for model robustness analysis, since its independence of input reveals the intrinsic characteristics of the model. Relatively, another interesting observation is Neural Collapse (NC), which means the feature variability may collapse during the terminal phase of training. Motivated by this, we propose to generate UAP by attacking the layer where NC phenomenon happens. Because of NC, the proposed attack could gather all the natural images’ features to its surrounding, which is hence called Feature-Gathering UAP (FG-UAP). We evaluate the effectiveness our proposed algorithm on abundant experiments, including untargeted and targeted universal attacks, attacks under limited dataset, and transfer-based black-box attacks among different architectures including Vision Transformers, which are believed to be more robust. After that, we empirically verify the effectiveness of NC's conclusion on UAP by attacking on only 10% of the dataset while keeping comparable performance. Finally, we investigate FG-UAP in the view of NC by analyzing the labels and extracted features of adversarial examples, finding that collapse phenomenon becomes stronger after the model is corrupted. Codes for the project are available at https://***/yzx1213/FG-UAP.

关键词：

SLoRD: Structural Low-Rank Descriptors for Shape Consistency in Vertebrae Segmentation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： You, Xin Lou, Yixin Zhang, Minghui Yang, Jie Navab, Nassir Gu, Yun Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai200240 China Technical University of Munich Munich Germany

Automatic and precise multi-class vertebrae segmentation from CT images is crucial for various clinical applications. However, due to a lack of explicit consistency constraints, existing methods especially for single-stage methods, still suffer from the challenge of intra-vertebrae segmentation inconsistency, which refers to multiple label predictions inside a singular vertebra. For multi-stage methods, vertebrae detection serving as the first step, tends to be affected by the pathology and metal implants. Thus, imprecise detections cause biased patches before segmentation, which then leads to inaccurate contour delineation and inconsistent segmentation. In our work, we intend to label individual and complete binary masks to address that challenge. Specifically, a contour generation network is proposed based on Structural Low-Rank Descriptors for shape consistency, termed SLoRD. For a structural representation of vertebral contours, we adopt the spherical coordinate system and devise the spherical centroid to calculate contour descriptors. Due to vertebrae’s similar appearances, basic contour descriptors can be acquired to restore original contours. Therefore, SLoRD leverages these contour priors and explicit shape constraints to facilitate regressed contour points close to vertebral surfaces. Quantitative and qualitative evaluations on VerSe 2019 and 2020 demonstrate the superior performance of our framework over other single-stage and multi-stage state-of-the-art (SOTA) methods. Further, SLoRD is a plug-and-play framework to refine the segmentation inconsistency existing in coarse predictions from other approaches. Source codes are available https://***/AlexYouXin/ SLoRD-VerSe. Copyright © 2024, The Authors. All rights reserved.

关键词： Metal implants

AttSiOff: a self-attention-based approach on siRNA design with inhibition and off-target effect prediction

学校读者我要写书评

暂无评论

Med-X 2024年第1期2卷 59-69页

作者： Bin Liu Ye Yuan Xiaoyong Pan Hong-Bin Shen Cheng Jin Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong UniversityShanghaiChina Medical Robot Research Institute School of Biomedical EngineeringShanghai Jiao Tong UniversityShanghaiChina

Small interfering RNA(siRNA)is often used for function study and expression regulation of specific genes,as well as the development of small molecule *** siRNAs with high inhibition and low off-target effects from massive can-didates is always a great *** experimentally-validated samples can prompt the development of machine-learning-based algorithms,including Support Vector Machine(SVM),Convolutional Neural Network(CNN),and Graph Neural Network(GNN).However,these methods still suffer from limited accuracy and poor generalization in designing potent and specific *** this study,we propose a novel approach for siRNA inhibition and off-target effect prediction,named *** combines a self-attention-based siRNA inhibition predictor with an mRNA searching package and an off-target *** predictor gives the inhibition score via analyzing the embedding of siRNA and local mRNA sequences,generated from the pre-trained RNA-FM model,as well as other meaningful prior-knowledge-based ***-attention mechanism can detect potentially decisive features,which may determine the inhibition of *** captures global and local dependencies more efficiently than normal *** tenfold cross-validation results indicate that our model outperforms all existing methods,achieving PCC of 0.81,SPCC of 0.84,and AUC of *** also reaches better performance of generalization and robustness on cross-dataset *** addition,the mRNA searching package could find all mature mRNAs for a given gene name from the GENOMES database,and the off-target filter can calculate the amount of unwanted off-target binding sites,which affects the specificity of *** on five mature siRNA drugs,as well as a new target gene(AGT),show that AttSioff has excellent convenience and operability in practical applications.

关键词： siRNA inhibition prediction Off-target effects Self-attention mechanism

Ocean archaea PPI prediction with pretraining models 25

学校读者我要写书评

暂无评论

Ocean archaea PPI prediction with pretraining models

2025 5th International Conference on Bioinformatics and Intelligent Computing, BIC 2025

作者： Zhang, Ying Liu, Yuan Pan, Xiaoyong Shen, Hongbin Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China

ISBN: (纸本)9798400712203

Protein-Protein Interaction (PPI) provides important insights into the metabolic mechanisms of different biological processes. Although PPIs in some organisms have been investigated systematically, PPIs in the ocean archaea remain largely unexplored. But such species have special investigation value since their adaptation to extreme living conditions may generate unique PPIs. In this paper, we aim to characterize and predict PPIs in ocean archaea to advance understanding of their metabolic networks. First, we collect all ocean archaea PPIs with high confidence from STRING database and analyze the PPI network features, including centrality and enrichment analysis. The functional enrichment results of the largest connecting subgraph in the PPI network show most PPIs in our constructed dataset is related to the translation and transcription processes. Then, we generate an equal number of negative PPI pairs, whose members have either different subcellular locations or GO terms. We also use the generated dataset to test the performance of three pretraining methods and their ensemble methods in the binary PPI prediction task. Our results suggest the ensemble methods could be applied to further improve models' performance. Fine-tuned models trained on the ocean archaea dataset are expected to predict the other ocean archaea PPIs that are not included in the STRING database and get more understanding about the ocean archaea PPI universe. © 2025 Copyright held by the owner/author(s).

关键词： Binary PPI prediction Ocean archaea PPI network Pretraining model

FASTC: A Fast Attentional Framework for Semantic Traversability Classification Using Point Cloud

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Yirui Wei, Pengjin Liu, Zhenhuan Wang, Bingchao Yang, Jie Liu, Wei Institute of Image Processing and Pattern Recognition Department of Automation Shanghai Jiao Tong University China Key Laboratory of System Control and Information Processing Ministry of Education China

Producing traversability maps and understanding the surroundings are crucial prerequisites for autonomous navigation. In this paper, we address the problem of traversability assessment using point clouds. We propose a novel pillar feature extraction module that utilizes PointNet to capture features from point clouds organized in vertical volume and a 2D encoder-decoder structure to conduct traversability classification instead of the widely used 3D convolutions. This results in less computational cost while even better performance is achieved at the same time. We then propose a new spatio-temporal attention module to fuse multi-frame information, which can properly handle the varying density problem of LIDAR point clouds, and this makes our module able to assess distant areas more accurately. Comprehensive experimental results on augmented Semantic KITTI and RELLIS-3D datasets show that our method is able to achieve superior performance over existing approaches both quantitatively and quantitatively. Our code is publicly available at https://***/chenyirui/FASTC. © 2024, CC BY.

关键词： Semantics

ProtDAT: A Unified Framework for Protein Sequence Design from Any Protein Text Description

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Guo, Xiao-Yu Li, Yi-Fan Liu, Yuan Pan, Xiaoyong Shen, Hong-Bin Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China

Protein design has become a critical method in advancing significant potential for various applications such as drug development and enzyme engineering. However, protein design methods utilizing large language models with solely pretraining and fine-tuning struggle to capture relationships in multi-modal protein data. To address this, we propose ProtDAT, a de novo fine-grained framework capable of designing proteins from any descriptive protein text input. ProtDAT builds upon the inherent characteristics of protein data to unify sequences and text as a cohesive whole rather than separate entities. It leverages an innovative multi-modal cross-attention, integrating protein sequences and textual information for a foundational level and seamless integration. Experimental results demonstrate that ProtDAT achieves the state-of-the-art performance in protein sequence generation, excelling in rationality, functionality, structural similarity, and validity. On 20,000 text-sequence pairs from Swiss-Prot, it improves pLDDT by 6%, TM-score by 0.26, and reduces RMSD by 1.2 Å, highlighting its potential to advance protein design. © 2024, CC BY-NC-SA.

关键词：

Quaternion CNN With Salient Features for Color image Denoising

学校读者我要写书评

暂无评论

Quaternion CNN With Salient Features for Color Image Denoisi...

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Yi Liu Qiyu Jin Jie Yang School of Mathematical Science Inner Mongolia University Hohhot China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Deep convolutional neural networks have significantly advanced color image denoising. However, existing models often apply grayscale denoising techniques to color images without accounting for inter-channel correlations, resulting in color distortion, detail loss, and visual artifacts. Moreover, these models frequently neglect salient features within convolutional maps. To address these issues, we propose a quaternion CNN model that captures channel correlations and extracts salient features, thereby enhancing color image denoising performance. Specifically, we convert color images into quaternion matrices to better capture these correlations and design a quaternion convolutional network to learn relevant features. Furthermore, an aggregated feature block is introduced to enhance the extraction of salient features and further refine the denoising process. Experimental results on multiple datasets demonstrate that the proposed model achieves superior performance compared to recent state-of-the-art methods.

关键词： Visualization Correlation Acoustic distortion image color analysis Convolution Quaternions Noise reduction Color Feature extraction Convolutional neural networks

Intergrated Segmentation and Detection Models for Dentex Challenge 2023

学校读者我要写书评

暂无评论

arXiv 2023年

作者： He, Lanshan Liu, Yusheng Wang, Lisheng Institute of Image Processing and Pattern Recognition Department of Automation Shanghai Jiao Tong University Shanghai200240 China

Dental panoramic x-rays are commonly used in dental diagnosing. With the development of deep learning, auto detection of diseases from dental panoramic x-rays can help dentists to diagnose diseases more efficiently. The Dentex Challenge 2023 is a competition for automatic detection of abnormal teeth along with their enumeration ids from dental panoramic x-rays. In this paper, we propose a method integrating segmentation and detection models to detect abnormal teeth as well as obtain their enumeration ids. Our codes are available at https://***/xyzlancehe/DentexSegAndDet. Copyright © 2023, The Authors. All rights reserved.

关键词： Diagnosis

Inferior Alveolar Nerve Segmentation in CBCT images using Connectivity-Based Selective Re-training

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Liu, Yusheng Xin, Rui Yang, Tao Wang, Lisheng Institute of Image Processing and Pattern Recognition Department of Automation Shanghai Jiao Tong University Shanghai200240 China

Inferior Alveolar Nerve (IAN) canal detection in CBCT is an important step in many dental and maxillofacial surgery applications to prevent irreversible damage to the nerve during the procedure. The ToothFairy2023 Challenge aims to establish a 3D maxillofacial dataset consisting of all sparse labels and partial dense labels, and improve the ability of automatic IAN segmentation. In this work, in order to avoid the negative impact brought by sparse labeling, we transform the mixed supervised problem into a semi-supervised problem. Inspired by self-training via pseudo labeling, we propose a selective re-training framework based on IAN connectivity. Our method is quantitatively evaluated on the ToothFairy verification cases, achieving the dice similarity coefficient (DSC) of 0.7956, and 95% hausdorff distance (HD95) of 4.4905, and wining the champion in the competition. Code is available at https://***/GaryNico517/SSLIAN-Retraining. © 2023, CC BY-NC-ND.

关键词： image segmentation