检索结果-内蒙古大学图书馆

9th International Conference on Computing and Artificial Intelligence, ICCAI 2023

作者： Li, Yanjun Ling, Xufeng Yang, Jie Shanghai Normal University Tianhua College AI School No. 1661 North Sheng Xin Road China Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University 800 Dongchuan Road 200240 China

ISBN: (纸本)9781450399029

Distracted driving is a phenomenon that drivers’ attention points to activities unrelated to normal driving, which leads to the decline of driving operation ability. With the popularity of smart phones, traffic accidents caused by distracted driving are on the rise. Traffic accidents caused by driver’s inattention due to vision deviation or distraction are particularly obvious in rear end collisions. The results of the distracted driving questionnaire show that almost all drivers have experienced distracted driving. In order to detect whether drivers are distracted when driving, this paper proposes a distracted driving analysis based on deep learning of human posture. This method uses the MMDetection framework in depth learning to build a target detection network to achieve target detection, then uses the depth neural network to estimate the geometric transformation features, and uses the stacked hourglass network to accurately predict the state features of drivers in the graph. The experimental results show that the measurement method has good accuracy, fast speed, good robustness and consistency, more accurate extraction of human posture related features, strong generalization ability and certain practical value. © 2023 Copyright held by the owner/author(s).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

An Automatic Key-points Detection and Style Transfer based Method of Articulatory Animation Generation 23

An Automatic Key-points Detection and Style Transfer based M...

引用

9th International Conference on Computing and Artificial Intelligence, ICCAI 2023

作者： Liu, Wei Ling, Xufeng Liang, Jingxin Wang, Xiaobo Yang, Jie Shanghai Normal University Tianhua College AI School No. 1661 North Sheng Xin Road China Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University 800 Dongchuan Road 200240 China

ISBN: (纸本)9781450399029

The method of generating articulatory animation based on automatic key point detection and style transfer. This paper proposes a system of key-points detection and registration based on image feature space, and uses the GAN for motion transfer, which is used to automatically generate oral and tongue articulatory animation of English. First, by obtaining the standard articulation image of the International Phonetic Alphabet from the public dataset, we train a deep-learning motion model according to the characteristics of the tongue, upper jaw, lower jaw, small tongue, soft palate and other motion features, according to the characteristics of articulatory organ, and automatically detect the key-points in the image feature space. According to the existing MRI articulatory videos, using different styles of articulatory images, a generative model is used to generate end-to-end animation of different styles automatically. For different combinations of consonants and vowels, generate animation of different syllables, and according to different categories of articulatory, such as stress and light tone, adjust the duration of articulatory to form different animation of stress and light tone of the same syllable. And form the animation of words through syllables, and then form the animation of each sentence. Experiments show that this method can simulate realistic articulatory animation according to input phonetic symbols, and has a very positive effect on English articulatory correction. © 2023 Copyright held by the owner/author(s).

关键词： Animation

来源：评论

学校读者我要写书评

暂无评论

Supervised Classification of Plant image Based on Attention Mechanism 7

Supervised Classification of Plant Image Based on Attention ...

引用

7th International Conference on Systems and Informatics, ICSAI 2021

作者： Li, Jie Yang, Jie Shanghai Jiao Tong University Institute of Image Processing and Pattern Recognition Shanghai China

ISBN: (纸本)9781665426244

In view of the wide variety of plants on the earth, the plant species identification is particularly necessary to protect and preserve biodiversity. In this work, we propose a plant image classification method based on the encoder-decoder model with additive attention mechanism to extract plant image features and convert them into text descriptions related to plant features. In a well-trained network, it can successfully classify on the species of the generated plant texts. We show that, the proposed method not only equalizes the results of deep convolutional neural network on classification task, but also uses of the prior information of botanists in classification, and thus provide a significant prediction result. © 2021 IEEE.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

Progressive low/high-resolution space attention fusion network for single image super-resolution

Progressive low/high-resolution space attention fusion netwo...

引用

2020 International Symposium on Automation, Information and Computing, ISAIC 2020

作者： Zhong, Chengzu Zhou, Yue Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China

The general single image super-resolution methods mainly extract features from the high-resolution (HR) space by the pre-upscaling step at the beginning of the network or from the low-resolution (LR) space before the post-upscaling step at the end of the network. The former way requires high computation as well as misleading the network by wrong artificial priors. The latter way cannot learn mapping well by only conducting simple operations in HR space. In this paper, we aim to utilize the features from LR and HR space more efficiently and propose the novel network, which applies a frequency-slicing mechanism to divide features into LR and HR space, a direction-aware fusion residual group to extract distinctive features in LR space and an attention fusion module to recalibrate features in HR space. The experimental results demonstrate that our model is superior to the state-of-the-art methods upon quantitative metrics and visual quality. © 2021 institute of Physics Publishing. All rights reserved.

关键词： Optical resolving power

来源：评论

学校读者我要写书评

暂无评论

Robot Debater: Debate-styled Text Auto-generation System Based on Large Foundation Language Models 4

Robot Debater: Debate-styled Text Auto-generation System Bas...

引用

4th IEEE International Conference on pattern recognition and Machine Learning, PRML 2023

作者： Zhu, Yu Ling, Yijun Ling, Xufeng Yang, Jie Shanghai China Shanghai Xiangming High School Shanghai China Shanghai Normal University Tianhua College School of Artificial Intelligence Shanghai China Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

ISBN: (纸本)9798350324303

We use a large foundation language model, which is fine-tuned with debate corpora, to develop a robot debater application. To address the limitations of requiring immense computational power in large base language models, this study takes advantage of the Low Rank Adaption characteristic prevalent in domain expert knowledge. By applying Low Rank Adaption technology and fine-tuning with a dedicated dataset, the computational load is reduced to just one-thousandth of what is needed for a large language model, greatly expanding the application scenarios of robot debaters using large language models. In view of the characteristics of debate competitions, this model can preset a variety of debate scenarios and supports personalized debate processes. It employs intelligent voice recognition technology combined with a multi-channel voice input method, allowing for precise localization of different human debaters and improving the accuracy of voice input recognition. The system can support multiple large-scale language generation models and utilize various different voice broadcasting systems, including male and female voice styles, as well as a range of voice emotions. This model can be applied to debate competitions held in universities, high schools, and various industries. It can support human-machine debates as well as machine-to-machine debates. © 2023 IEEE.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Towards Better Data Exploitation in Self-Supervised Monocular Depth Estimation

arXiv

引用

arXiv 2023年

作者： Liu, Jinfeng Kong, Lingtong Yang, Jie Liu, Wei Institute of Image Processing and Pattern Recognition Department of Automation Shanghai Jiao Tong University Shanghai China

Depth estimation plays an important role in robotic perception systems. The self-supervised monocular paradigm has gained significant attention since it can free training from the reliance on depth annotations. Despite recent advancements, existing self-supervised methods still underutilize the available training data, limiting their generalization ability. In this paper, we take two data augmentation techniques, namely Resizing-Cropping and Splitting-Permuting, to fully exploit the potential of training datasets. Specifically, the original image and the generated two augmented images are fed into the training pipeline simultaneously and we leverage them to conduct self-distillation. Additionally, we introduce the detail-enhanced DepthNet with an extra full-scale branch in the encoder and a grid decoder to enhance the restoration of fine details in depth maps. Experimental results demonstrate our method can achieve state-of-the-art performance on the KITTI and Cityscapes datasets. Moreover, our KITTI models also show superior generalization performance when transferring to Make3D, NYUv2 and Cityscapes datasets. Our codes are available at https://***/LiuJF1226/BDEdepth. Copyright © 2023, The Authors. All rights reserved.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

MUSO: Achieving Exact Machine Unlearning in Over-Parameterized Regimes

arXiv

引用

arXiv 2024年

作者： Yang, Ruikai He, Mingzhen He, Zhenghao Qiu, Youmei Huang, Xiaolin Institute of Image Processing and Pattern Recognition Department of Automation Shanghai Jiao Tong University 800 Dongchuan RD Shanghai200240 China

Machine unlearning (MU) is to make a well-trained model behave as if it had never been trained on specific data. In today’s over-parameterized models, dominated by neural networks, a common approach is to manually relabel data and fine-tune the well-trained model. It can approximate the MU model in the output space, but the question remains whether it can achieve exact MU, i.e., in the parameter space. We answer this question by employing random feature techniques to construct an analytical framework. Under the premise of model optimization via stochastic gradient descent, we theoretically demonstrated that over-parameterized linear models can achieve exact MU through relabeling specific data. We also extend this work to real-world nonlinear networks and propose an alternating optimization algorithm that unifies the tasks of unlearning and relabeling. The algorithm’s effectiveness, confirmed through numerical experiments, highlights its superior performance in unlearning across various scenarios compared to current state-of-the-art methods, particularly excelling over similar relabeling-based MU approaches. Copyright © 2024, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

FG-UAP: Feature-Gathering Universal Adversarial Perturbation

FG-UAP: Feature-Gathering Universal Adversarial Perturbation

引用

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：

来源：评论

学校读者我要写书评

暂无评论

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

arXiv

引用

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

来源：评论

学校读者我要写书评

暂无评论

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

引用

Med-X 2024年第1期2卷 1-11页

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

Small interfering RNA (siRNA) is often used for function study and expression regulation of specific genes, as well as the development of small molecule drugs. Selecting siRNAs with high inhibition and low off-target effects from massive candidates is always a great challenge. Increasing 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 siRNAs. In this study, we propose a novel approach for siRNA inhibition and off-target effect prediction, named AttSiOff. It combines a self-attention-based siRNA inhibition predictor with an mRNA searching package and an off-target filter. The 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 features. Self-attention mechanism can detect potentially decisive features, which may determine the inhibition of siRNA. It captures global and local dependencies more efficiently than normal convolutions. The tenfold cross-validation results indicate that our model outperforms all existing methods, achieving PCC of 0.81, SPCC of 0.84, and AUC of 0.886. It also reaches better performance of generalization and robustness on cross-dataset validation. In 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 siRNA. Experiments on five mature siRNA drugs, as well as a new target gene (AGT), show that AttSioff has excellent convenience and operability in practical applications. · We utilize a pre-trained model to enrich the information of sequence embedding,

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：