检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Guo, Pinxue Li, Wanyun Huang, Hao Hong, Lingyi Zhou, Xinyu Chen, Zhaoyu Li, Jinglun Jiang, Kaixun Zhang, Wei Zhang, Wenqiang Shanghai Engineering Research Center of AI & Robotics Academy for Engineering & Technology Fudan University China Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University China Engineering Research Center of AI & Robotics Ministry of Education Academy for Engineering & Technology Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University China

Multi-modal Video Object Segmentation (VOS), including RGB-Thermal, RGB-Depth, and RGB-Event, has garnered attention due to its capability to address challenging scenarios where traditional VOS methods struggle, such as extreme illumination, rapid motion, and background distraction. Existing approaches often involve designing specific additional branches and performing full-parameter fine-tuning for fusion in each task. However, this paradigm not only duplicates research efforts and hardware costs but also risks model collapse with the limited multi-modal annotated data. In this paper, we propose a universal framework named X-Prompt for all multi-modal video object segmentation tasks, designated as RGB+X. The X-Prompt framework first pretrains a video object segmentation foundation model using RGB data, and then utilize the additional modality of the prompt to adapt it to downstream multi-modal tasks with limited data. Within the X-Prompt framework, we introduce the Multi-modal Visual Prompter (MVP), which allows prompting foundation model with the various modalities to segment objects precisely. We further propose the Multi-modal Adaptation Experts (MAEs) to adapt the foundation model with pluggable modality-specific knowledge without compromising the generalization capacity. To evaluate the effectiveness of the X-Prompt framework, we conduct extensive experiments on 3 tasks across 4 benchmarks. The proposed universal X-Prompt framework consistently outperforms the full fine-tuning paradigm and achieves state-of-the-art performance. Code: https://***/PinxueGuo/***. Copyright © 2024, The Authors. All rights reserved.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

EC-SLAM: Effectively Constrained Neural RGB-D SLAM with Sparse TSDF Encoding and Global Bundle Adjustment

arXiv

引用

arXiv 2024年

作者： Li, Guanghao Chen, Qi Yan, Yuxiang Pu, Jian Institute of Science and Technology for Brain-Inspired Intelligence Fudan University Shanghai200433 China Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University Shanghai200433 China

We introduce EC-SLAM, a real-time dense RGB-D simultaneous localization and mapping (SLAM) system leveraging Neural Radiance Fields (NeRF). While recent NeRF-based SLAM systems have shown promising results, they have yet to fully exploit NeRF’s potential to constrain pose optimization. EC-SLAM addresses this by using sparse parametric encodings and Truncated Signed Distance Fields (TSDF) to represent the map, enabling efficient fusion, reducing model parameters, and accelerating convergence. Our system also employs a globally constrained Bundle Adjustment (BA) strategy that capitalizes on NeRF’s implicit loop closure correction capability, improving tracking accuracy by reinforcing constraints on keyframes most relevant to the current optimized frame. Furthermore, by integrating a feature-based and uniform sampling strategy that minimizes ineffective constraint points for pose optimization, we reduce the impact of random sampling in NeRF. Extensive evaluations on the Replica, ScanNet, and TUM datasets demonstrate state-of-the-art performance, with precise tracking and reconstruction accuracy achieved alongside real-time operation at up to 21 Hz. The source code is available at https://***/Lightingooo/EC-SLAM. Copyright © 2024, The Authors. All rights reserved.

关键词： Mapping

来源：评论

学校读者我要写书评

暂无评论

Boosting Fine-Grained Sketch-Based Image Retrieval with Self-Supervised Learning

Boosting Fine-Grained Sketch-Based Image Retrieval with Self...

引用

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

作者： Zhaolong Zhang Yangdong Chen Yuejie Zhang Rui Feng Tao Zhang Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University Shanghai China School of Information Management and Engineering Shanghai University of Finance and Economics Shanghai China

Fine-grained sketch-based image retrieval (FG-SBIR) aims at aligning images and sketches at the instance level. It is a challenging task as there are significant differences between sketch and image. Existing methods usually produce less desired performance due to the lack of large-scale fine-grained image-sketch datasets and the strong dependence on the classification models pretrained on ImageNet. In this paper, we propose a better self-supervised pre-trained FG-SBIR model which does not depend on large-scale annotated datasets. Only images and their corresponding edge maps are used at the pre-training stage. Mixed modal transformation is designed to generate different mixed-up views. The FG-SBIR model is pre-trained by minimizing the distance between the views of the same instance and then fine-tuned by a simple triplet loss. With a plain downstream network, it achieves generally better performance than state-of-the-art models on three widely used FG-SBIR datasets.

关键词： Image edge detection Image retrieval Self-supervised learning Signal processing Boosting Acoustics Task analysis

来源：评论

学校读者我要写书评

暂无评论

Minimizing Disparities between Real and Pseudo Queries for Unsupervised Visual Grounding

Minimizing Disparities between Real and Pseudo Queries for U...

引用

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

作者： Hui Jiang Changkai Ji Jilan Xu Yanhao Zhu Yuejie Zhang Rui Feng Tao Zhang Shang Gao School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Fudan University Shanghai China School of Information Management and Engineering Shanghai Key Laboratory of Financial Information Technology Shanghai University of Finance and Economics Shanghai China School of Information Technology Deakin University Victoria Australia

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Visual grounding involves the identification and localization of image regions given textual descriptions. To reduce the manual labeling effort on region-text pairs, unsupervised visual grounding aims to generate pseudo bounding box and query pairs for training grounding models. However, there exists significant disparities between real and pseudo queries in terms of object, attribute distributions, and textual formats, limiting the generalization performance of unsupervised grounding methods. To address this challenge, we propose a novel unsupervised visual grounding framework. During training, we prompt Multimodal Large Language Models to generate pseudo queries, in which the entities are beyond the object detector’s pre-defined limited categories, and are associated with richer attributes. We further devise a Modifier Tree structure to bridge the gap of textual format between real and pseudo queries. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art unsupervised approaches on public benchmark datasets, particularly when dealing with complex queries.

关键词： Training Location awareness Visualization Limiting Grounding Large language models Manuals labeling Speech processing Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Read Extensively, Focus Smartly: A Cross-document Semantic Enhancement Method for Visual Documents NER 29

Read Extensively, Focus Smartly: A Cross-document Semantic E...

引用

29th International Conference on Computational Linguistics, COLING 2022

作者： Zhao, Jun Zhao, Xin Zhan, Wenyu Gui, Tao Zhang, Qi Qiao, Liang Cheng, Zhanzhan Pu, Shiliang School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Fudan University Shanghai China Institute of Modern Languages and Linguistics Fudan University China Hikvision Research Institute Shanghai China

The introduction of multimodal information and pretraining technique significantly improves entity recognition from visually-rich documents. However, most of the existing methods pay unnecessary attention to irrelevant regions of the current document while ignoring the potentially valuable information in related documents. To deal with this problem, this work proposes a cross-document semantic enhancement method, which consists of two modules: 1) To prevent distractions from irrelevant regions in the current document, we design a learnable attention mask mechanism, which is used to adaptively filter redundant information in the current document. 2) To further enrich the entity-related context, we propose a cross-document information awareness technique, which enables the model to collect more evidence across documents to assist in prediction. The experimental results on two documents understanding benchmarks covering eight languages demonstrate that our method outperforms the SOTA methods. © 2022 Proceedings - International Conference on Computational Linguistics, COLING. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Competition on robust deep learning

引用

National science Review 2023年第6期10卷 13-15页

作者： Yinpeng Dong Chang Liu Wenzhao Xiang Hang Su Jun Zhu Department of Computer Science and Technology Institute for AI Tsinghua-Bosch Joint ML Center THBI Lab BNRist Center Tsinghua University Institute of Image Communication and Networks Engineering in the Department of Electronic Engineering (EE) Shanghai Jiao Tong University Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Institute of Computing Technology CAS Peng Cheng Laboratory Pazhou Laboratory (Huangpu)

PROBLEM In recent years,the rapid development of artificial intelligence (AI) technology,especially machine learning and deep learning, is profoundly changing human production and *** various fields,such as robotics,face recognition,autonomous driving and healthcare,AI is playing an important ***,although AI is promoting the technological revolution and industrial progress,its security risks are often *** studies have found that the wellperforming deep learning models are extremely vulnerable to adversarial examples [1-3].The adversarial examples are crafted by applying small,humanimperceptible perturbations to natural examples,but can mislead deep learning models to make wrong *** vulnerability of deep learning models to adversarial examples can raise security and safety threats to various realworld applications.

关键词： robustness adversarial example adversarial training deep learning

来源：评论

学校读者我要写书评

暂无评论

Pseudo-3D CNN with inter-slice attention for glioma grading 3

Pseudo-3D CNN with inter-slice attention for glioma grading

引用

3rd International Conference on computer science and Communication Technology, ICCSCT 2022

作者： Du, Shanshan Cao, Zhuoqun Zhao, Rui-Wei Zhang, Xiaobo Feng, Rui School of Information Science and Technology Fudan University Shanghai China School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Fudan University Shanghai China Academy for Engineering and Technology Fudan University Shanghai China Children's Hospital Fudan University Shanghai China

ISBN: (纸本)9781510661240

Gliomas comprise around 80 percent of all malignant brain tumors which can be further classified into the low-grade glioma and high-grade glioma categories. Compared to low-grade gliomas, high-grade gliomas show more malignant behaviour since they usually grow rapidly and frequently destroy healthy brain tissue. Therefore, it is important to determine the malignancy of gliomas for initial treatment plan. Most existing methods are based on radiomics or transfer learning which extract features from single-slice without considering that 3D features between adjacent slices can provide stronger discriminative power. In this work, we propose to incorporate the attention mechanism and pseudo-3d module into a deep convolutional network architecture, which can reduce model size and produce fine features for glioma grading. Our work focuses on the inter-slice relationship and propose an attention unit, named "Inter-Slice Attention Module", which adaptively refines intermediate feature maps by modelling dependencies between adjacent slices. We evaluate our method on an open dataset of gliomas, achieving mean accuracy of 89.47%. © 2022 SPIE.

关键词： Grading

来源：评论

学校读者我要写书评

暂无评论

Gaitcotr: Improved Spatial-Temporal Representation for Gait Recognition with a Hybrid Convolution-Transformer Framework

Gaitcotr: Improved Spatial-Temporal Representation for Gait ...

引用

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

作者： Jingqi Li Yuzhen Zhang Hongming Shan Junping Zhang Shanghai Key Lab of Intelligent Information Processing School of Computer Science Institute of Science and Technology for Brain-Inspired Intelligence Fudan University Shanghai China

This work presents a novel hybrid convolution-transformer framework for gait recognition, termed GaitCoTr. The developed framework captures the appearance and short-term temporal features by convolution and extracts the long-term temporal features by transformer architecture, achieving a comprehensive spatial-temporal representation of gait. To unleash the potential of this hybrid framework for extracting richness and generalized temporal features, we propose a new variant of transformer tailored for gait, including temporally shifted tokenization, length-flexible position embedding, and inter-frame encoder. In addition, we introduce an auxiliary task—view label prediction—aiming to disentangle view from ID information. Extensive experimental results on two well-known gait benchmark datasets, CASIA-B and GREW, demonstrate the superior performance of the proposed Gait-CoTr.

关键词： Convolution Speech recognition Benchmark testing Feature extraction Transformers Tokenization Acoustics

来源：评论

学校读者我要写书评

暂无评论

Local to global Tracker: A Siamese Network for Long-term Tracking 3

Local to global Tracker: A Siamese Network for Long-term Tra...

引用

2021 3rd International Conference on Robotics, intelligent Control and Artificial Intelligence, ICRICA 2021

作者： Yang, Ya Zhou, Xinyu Jiang, Kaixun Hou, Xiaonan Wang, Chunlei Zhang, Wenqiang 21st Research Institute of China Electronics Technology Group Corporation Shanghai China Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Fudan University Shanghai China

Visual object tracking is a basic task in computer vision, which can be applied to different applications. Former methods of single object tracking problem demands to perform target state estimation according to object position of previous frame, but these methods are poorly capable of handling the fast movement of the target. In this paper, we introduce a Siamese network for single object tracking task. The network consists of two branches, one of which is the classification branch used to predict positive or negative samples, and the other is the regression branch used to predict the specific location of the object in sequence. In addition, we propose a local to global algorithm for long-term tracking to expand the search region when the target is out-of-view or full occlusion. Experiments show that our tracker achieves very promising results on VOT2018LT datasets while running at over 89 FPS. © Published under licence by IOP Publishing Ltd.

关键词： Target tracking

来源：评论

学校读者我要写书评

暂无评论

贝叶斯推理和动态神经反馈促进先天性心脏病智能诊断的临床应用

引用

Engineering 2023年第4期23卷 90-102,M0005页

作者：谭伟敏曹银银马晓静茹港徽李吉春张璟高燕杨佳伦黄国英颜波李健 Shanghai Key Laboratory of Intelligent Information Processing School of Computer ScienceFudan UniversityShanghai 200433China Cardiovascular Center&Clinical Laboratory Center Children’s Hospital of Fudan UniversityNational Children’s Medical CenterShanghai 201102China

先天性心脏病(CHD)是婴幼儿死亡主要原因。基于人工智能的先天性心脏病诊断网络(CHDNet)是一种基于超声心动图视频的二分类模型,用于判别超声心动图视频是否包含心脏缺陷。现有的CHDNet模型表现出与医学专家相当甚至更好的判别性能,但... 详细信息

先天性心脏病(CHD)是婴幼儿死亡主要原因。基于人工智能的先天性心脏病诊断网络(CHDNet)是一种基于超声心动图视频的二分类模型,用于判别超声心动图视频是否包含心脏缺陷。现有的CHDNet模型表现出与医学专家相当甚至更好的判别性能,但它们在训练集之外的样本上的不可靠性已成为模型部署的关键瓶颈。而这是当前大多数基于AI诊断方法的共性问题。为了克服这一挑战,本文我们提出了两种基本机制——贝叶斯推理和动态神经反馈——分别用于衡量和提高人工智能诊断的可靠性。贝叶斯推理允许神经网络模型输出CHD判别的可靠性而不仅仅是单一的判别结果,而动态神经反馈是一个计算神经反馈单元,允许神经网络将知识从输出层反馈给浅层,使神经网络有选择地激活相关神经元。为了评估这两种机制的有效性,我们在包含三种常见CHD缺陷的4151个超声心动图视频上训练了CHDNet,并在1037个超声心动图视频的内部测试集和从其他心血管成像设备新收集的692个外部视频集上对其进行了测试。每个超声心动图视频对应于一位患者和一次就诊。我们在多种代表性神经网络架构上展示了贝叶斯推理获得的可靠性如何解释和量化神经网络内部与外部测试集之间的性能显著差异,以及尽管输入被噪声破坏或使用外部测试集时,设计的反馈单元如何帮助神经网络保持高精度和可靠性。

关键词： Congenital heart disease Artificial intelligence Deep learning Model uncertainty

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：