检索结果-内蒙古大学图书馆

Automatic detection of the Pharyngeal Phase in Raw Videos for the Videofluoroscopic Swallowing Study Using Efficient data Collection and 3d convolutional networks

引用

SENSORS 2019年第18期19卷 3873-3873页

作者： Lee, Jong Taek Park, Eunhee Jung, Tae-du ETRI Artificial Intelligence Applicat Res Sect Daegu 42994 South Korea Kyungpook Natl Univ Chilgok Hosp Dept Rehabil Med Daegu 41404 South Korea Kyungpook Natl Univ Sch Med Dept Rehabil Med Daegu 41944 South Korea

Videofluoroscopic swallowing study (VFSS) is a standard diagnostic tool for dysphagia. To detect the presence of aspiration during a swallow, a manual search is commonly used to mark the time intervals of the pharyngeal phase on the corresponding VFSS image. In this study, we present a novel approach that uses 3d convolutional networks to detect the pharyngeal phase in raw VFSS videos without manual annotations. For efficient collection of training data, we propose a cascade framework which no longer requires time intervals of the swallowing process nor the manual marking of anatomical positions for detection. For video classification, we applied the inflated 3d convolutional network (I3d), one of the state-of-the-art network for action classification, as a baseline architecture. We also present a modified 3d convolutional network architecture that is derived from the baseline I3d architecture. The classification and detection performance of these two architectures were evaluated for comparison. The experimental results show that the proposed model outperformed the baseline I3d model in the condition where both models are trained with random weights. We conclude that the proposed method greatly reduces the examination time of the VFSS images with a low miss rate.

关键词： action detection action classification 3d convolutional networks pharyngeal phase videofluoroscopic swallowing study

来源：评论

学校读者我要写书评

暂无评论

deep Optical Flow Feature Fusion Based on 3d convolutional networks for Video Action Recognition 4

Deep Optical Flow Feature Fusion Based on 3D Convolutional N...

引用

IEEE SmartWorld, Ubiquitous Intelligence & Computing, Advanced & Trusted Computing, Scalable Computing & Communications, Cloud & Big data Computing, Internet of People and Smart City Innovation (SmartWorld/SCALCOM/UIC/ATC/CBdCom/IOP/SCI)

作者： Lu, Tongwei Ai, Shihui Jiang, Yongyuan Xiong, Yudian Min, Feng Wuhan Inst Technol Hubei Key Lab Intelligent Robot Coll Comp Sci & Engn Wuhan Hubei Peoples R China

ISBN: (纸本)9781538693803

In the real world, due to the existence of complex scenes and different perspectives, different types of behavior are very different in appearance and behavior models. The traditional 3d convolutional networks have greatly improved the extraction of time series information, but at the same time it also loses some behavioral characteristics, which lead to the recognition rate not high. And the optical flow information can represent the motion information of human behavior well, so extract more effective optical flow information before extracting RGB depth features by the traditional 3d convolutional networks is necessary. Extracting the deep RGB feature and the deep optical flow feature respectively by the feature extractor with 3d convolutional networks, Then the cascade fusion of two deep features make the feature stronger classification ability. The experimental results show that the improved recognition method has a better performance for the video behavior recognition than the traditional 3d convolutional networks method.

关键词： optical flow feature cascade fusion 3d convolutional networks libSVM

来源：评论

学校读者我要写书评

暂无评论

End-to-end video background subtraction with 3d convolutional neural networks

引用

MULTIMEdIA TOOLS ANd APPLICATIONS 2018年第17期77卷 23023-23041页

作者： Sakkos, dimitrios Liu, Heng Han, Jungong Shao, Ling Northumbria Univ Dept Comp & Informat Sci Newcastle Upon Tyne NE1 8ST Tyne & Wear England Anhui Univ Technol Sch Comp Sci & Technol Anhui Sheng 243032 Peoples R China Univ Lancaster Sch Comp & Commun Lancaster LA1 4YW England Univ East Anglia Sch Comp Sci Norwich NR4 7TJ Norfolk England

Background subtraction in videos is a highly challenging task by definition, as it lays on a pixel-wise classification level. Therefore, great attention to detail is essential. In this paper, we follow the success of deep Learning in Computer Vision and present an end-to-end system for background subtraction in videos. Our model is able to track temporal changes in a video sequence by applying 3d convolutions to the most recent frames of the video. Thus, no background model is needed to be retained and updated. In addition, it can handle multiple scenes without further fine-tuning on each scene individually. We evaluate our system on the largest dataset for change detection, Cdnet, with over 50 videos which span across 11 categories. Further evaluation is performed in the ESI dataset which features extreme and sudden illumination changes. Our model surpasses the state-of-the-art on both datasets according to the average ranking of the models over a wide range of metrics.

关键词： Computer vision deep learning Fully convolutional networks Background subtraction Video segmentation 3d convolutional networks

来源：评论

学校读者我要写书评

暂无评论

Soccer Video Event detection Using 3d convolutional networks and Shot Boundary detection via deep Feature distance 24th

Soccer Video Event Detection Using 3D Convolutional Networks...

引用

24th International Conference on Neural Information Processing (ICONIP)

作者： Liu, Tingxi Lu, Yao Lei, Xiaoyu Zhang, Lijing Wang, Haoyu Huang, Wei Wang, Zijian Beijing Inst Technol Sch Comp Sci Beijing Lab Intelligent Informat Technol Beijing Peoples R China China Cent Televis Beijing Peoples R China

ISBN: (纸本)9783319700960;9783319700953

In this work, we propose a novel framework combining temporal action localization and play-break (PB) rules for soccer video event detection. Firstly we treat event detection task in action-level, and adopt 3d convolutional networks to perform action localization. Then we employ PB rules to organize actions into events using long view and replay logo detected in the first step. Finally, we determine the semantic classes of events according to principal actions which contain key semantic information of highlights. For long untrimmed videos, we propose a shot boundary detection method using deep feature distance (dFd) to reduce the number of proposals and improve the performance of localization. Experiment results verify the effectiveness of our framework on a new dataset which contains 152 classes of semantic actions and scenes in soccer video.

关键词： Soccer event detection Temporal action localization 3d convolutional networks deep feature distance

来源：评论

学校读者我要写书评

暂无评论

depthwise Spatio-Temporal STFT Convolutiona Neural networks for Human Action Recognition

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS ANd MACHINE INTELLIGENCE 2022年第9期44卷 4839-4851页

作者： Kumawat, Sudhakar Verma, Manisha Nakashima, Yuta Raman, Shanmuganathan Osaka Univ Inst Databil Sci Intelligence & Sensing Lab Osaka 5670871 Japan Indian Inst Technol Gandhinagar Elect Engn & Comp Sci & Engn Gandhinagar 382355 Gujarat India

Conventional 3d convolutional neural networks (CNNs) are computationally expensive, memory intensive, prone to overfitting, and most importantly, there is a need to improve their feature learning capabilities. To address these issues, we propose spatio-temporal short-term Fourier transform (STFT) blocks, a new class of convolutional blocks that can serve as an alternative to the 3d convolutional layer and its variants in 3d CNNs. An STFT block consists of non-trainable convolution layers that capture spatially and/or temporally local Fourier information using an STFT kernel at multiple low frequency points, followed by a set of trainable linear weights for learning channel correlations. The STFT blocks significantly reduce the space-time complexity in 3d CNNs. In general, they use 3.5 to 4.5 times less parameters and 1.5 to 1.8 times less computational costs when compared to the state-of-the-art methods. Furthermore, their feature learning capabilities are significantly better than the conventional 3d convolutional layer and its variants. Our extensive evaluation on seven action recognition datasets, including Something(2) v1 and v2, Jester, diving-48, Kinetics-400, UCF 101, and HMdB 51, demonstrate that STFT blocks based 3d CNNs achieve on par or even better performance compared to the state-of-the-art methods.

关键词： Short-term fourier transform 3d convolutional networks human action recognition

来源：评论

学校读者我要写书评

暂无评论

A method for recognizing abnormal behaviors of personnel at petroleum stations based on GTB-ResNet

引用

International Journal of Intelligent Computing and Cybernetics 2024年第4期17卷 869-889页

作者： Huiling Yu Sijia dai Shen Shi Yizhuo Zhang Changzhou University ChangzhouChina School of Computer Science and Artificial Intelligence Changzhou UniversityChangzhouChina

Purpose-The abnormal behaviors of staff at petroleum stations pose significant safety *** the challenges of high parameter counts,lengthy training periods and low recognition rates in existing 3d ResNet behavior recognition models,this paper proposes GTB-ResNet,a network designed to detect abnormal behaviors in petroleum station ***/methodology/approach-Firstly,to mitigate the issues of excessive parameters and computational complexity in 3d ResNet,a lightweight residual convolution module called the Ghost residual module(GhostNet)is introduced in the feature extraction *** convolution replaces standard convolution,reducing model parameters while preserving multi-scale feature extraction ***,to enhance the model’s focus on salient features amidst wide surveillance ranges and small target objects,the triplet attention mechanism module is integrated to facilitate spatial and channel information ***,to address the challenge of short time-series features leading to misjudgments in similar actions,a bidirectional gated recurrent network is added to the feature extraction backbone *** ensures the extraction of key long time-series features,thereby improving feature extraction ***-The experimental setup encompasses four behavior types:illegal phone answering,smoking,falling(abnormal)and touching the face(normal),comprising a total of 892 *** results showcase GTB-ResNet achieving a recognition accuracy of 96.7%with a model parameter count of 4.46 M and a computational complexity of 3.898 *** represents a 4.4%improvement over 3d ResNet,with reductions of 90.4%in parameters and 61.5%in computational ***/value-Specifically designed for edge devices in oil stations,the 3d ResNet network is tailored for real-time action *** address the challenges posed by the large number of parameters in 3d ResNet networks and the difficulties in deployment o

关键词： Abnormal behavior recognition Lightweight residual convolutional GhostNet Triplet attention mechanism Bidirectional gated recurrent network Bi-GRU 3d convolutional networks

来源：评论

学校读者我要写书评

暂无评论

Basketball action recognition by fusing video recognition techniques with an SSd target detection algorithm

引用

OPEN COMPUTER SCIENCE 2025年第1期15卷

作者： He, Weizhao Li, Bin Guangzhou City Univ Technol Sports Dept Guangzhou 510800 Peoples R China

In the digital era, computer vision technology has become a key tool in the field of sports analysis, especially widely used in basketball action recognition (BAR) technology. To improve the recognition accuracy and efficiency of technical actions in basketball, a BAR system that integrates three-dimensional convolutional neural network video recognition technology and a single-shot multibox detector (SSd) target detection algorithm is proposed. First, the basketball player's action sequence is captured by the video recognition technique, and then the SSd target detection algorithm is utilized for real-time target detection and localization of the player's key parts. By combining these two techniques, it is possible to more accurately recognize and classify a wide range of basketball player's actions including but not limited to shoot, dribbling, passing, and defending. The experimental results indicated that the proposed model of the study provided the best recognition effect in continuous video images with 16 frames, as well as faster response and convergence speeds, and displayed good stability in the computational process. In the application effect comparison experiments with the comparison algorithms, the average recognition rate of the proposed model was 89.5 and 92.5% on the original and cropped frames, respectively, and the average accuracy mean value was 91.69%. Except for the shoot action, the recognition misjudgment rate for the remaining five basketball actions was the lowest among the three methods, which significantly improved the accuracy and effectiveness of action recognition. In addition, the accuracy of the proposed model in practical applications reached 94.50%, and the real-time performance reached 95 ms/frame, both significantly better than the other two compared models, indicating that the model is suitable for real-time BAR tasks and can provide efficient and accurate results.

关键词： basketball action recognition video recognition techniques SSd target detection 3d convolutional networks action classification

来源：评论

学校读者我要写书评

暂无评论

3dSleepNet: A Multi-Channel Bio-Signal Based Sleep Stages Classification Method Using deep Learning

引用

IEEE TRANSACTIONS ON NEURAL SYSTEMS ANd REHABILITATION ENGINEERING 2023年 31卷 3513-3523页

作者： Ji, Xiaopeng Li, Yan Wen, Peng Univ Southern Queensland Sch Math Phys & Comp Toowoomba Qld 4350 Australia Univ Southern Queensland Sch Engn Toowoomba Qld 4350 Australia

A novel multi-channel-based 3d convolutional neural network (3d-CNN) is proposed in this paper to classify sleep stages. Time domain features, frequency domain features, and time-frequency domain features are extracted from electroencephalography (EEG), electromyogram (EMG), and electrooculogram (EOG) channels and fed into the 3d-CNN model to classify sleep stages. Intrinsic connections among different bio-signals and different frequency bands in time series and time-frequency are learned by 3d convolutional layers, while the frequency relations are learned by 2d convolutional layers. Partial dot-product attention layers help this model find the most important channels and frequency bands in different sleep stages. A long short-term memory unit is added to learn the transition rules among neighboring epochs. Classification experiments were conducted using both ISRUC-S3 datasets and ISRUC-S1, sleep-disorder datasets. The experimental results showed that the overall accuracy achieved 0.832 and the F1-score and Cohen's kappa reached 0.814 and 0.783, respectively, on ISRUC-S3, which are a competitive classification performance with the state-of-the-art baselines. The overall accuracy, F1-score, and Cohen's kappa on ISRUC-S1 achieved 0.820, 0.797, and 0.768, respectively, which also demonstrate its generality on unhealthy subjects. Further experiments were conducted on ISRUC-S3 subset to evaluate its training time. The training time on 10 subjects from ISRUC-S3 with 8549 epochs is 4493s, which indicates its highest calculation speed compared with the existing high-performance graph convolutional networks and U-2 Net architecture algorithms.

关键词： Feature extraction Sleep Three-dimensional displays Time-frequency analysis Convolution convolutional neural networks Brain modeling deep learning 3d convolutional networks sleep stages classification

来源：评论

学校读者我要写书评

暂无评论

SE-Md: a single-encoder multiple-decoder deep network for point cloud reconstruction from 2d images

引用

PATTERN ANALYSIS ANd APPLICATIONS 2023年第3期26卷 1291-1302页

作者： Hafiz, Abdul Mueed Bhat, Rouf Ul Alam Parah, Shabir Ahmad Hassaballah, M. Univ Kashmir Dept Elect & Commun Engn Inst Technol Srinagar 190006 J&K India Univ Kashmir Dept Elect & Instrumentat Technol Srinagar 190006 J&K India Prince Sattam Bin Abdulaziz Univ Dept Comp Sci Coll Comp Engn & Sci Al Kharj 16278 Saudi Arabia South Valley Univ Fac Comp & Informat Dept Comp Sci Qena Egypt

3d model reconstruction from single 2d RGB images is a challenging and actively researched computer vision task. Several techniques based on conventional network architectures have been proposed for the same. However, the body of research work is limited and there are some issues like using inefficient 3d representation formats, weak 3d model reconstruction backbones, inability to reconstruct dense point clouds, dependence of post-processing for reconstruction of dense point clouds and dependence on silhouettes in RGB images. In this paper, a new 2d RGB image to point cloud conversion technique is proposed, which improves the state-of-the-art in the field due to its efficient, robust and simple model by using the concept of parallelization in network architecture. It not only uses efficient and rich 3d representation of point clouds, but also uses a new robust point cloud reconstruction backbone to address the prevalent issues. This involves using a single-encoder multiple-decoder deep network architecture wherein each decoder reconstructs certain fixed viewpoints. This is followed by fusing all the viewpoints to reconstruct a dense point cloud. Various experiments are conducted to evaluate the proposed technique and to compare its performance with those of the state-of-the-arts and impressive gains in performance are demonstrated.

关键词： 3d model reconstruction 3d shape reconstruction 2d images Point clouds ShapeNet 3d convolutional networks

来源：评论

学校读者我要写书评

暂无评论

A 3d attention networks for classification of white blood cells from microscopy hyperspectral images

引用

OPTICS ANd LASER TECHNOLOGY 2021年 139卷 106931-106931页

作者： Wang, Qian Wang, Jianbiao Zhou, Mei Li, Qingli Wen, Ying Chu, Junhao East China Normal Univ Shanghai Key Lab Multidimens Informat Proc Shanghai 200241 Peoples R China Engn Ctr SHMEC Space Informat Shanghai 200241 Peoples R China GNSS Shanghai 200241 Peoples R China East China Normal Univ Engn Res Ctr Nanophoton & Adv Instrument Minist Educ Shanghai 200241 Peoples R China Shanghai Jiao Tong Univ Med Sch Affiliated Ruijin Hosp Shanghai 200020 Peoples R China

Peripheral blood smear analysis is a key process for hematologists to reflect condition of human immune system and one of the earliest clinical applications that benefit from automatic computer-aided analysis. Changes in the ratio of white blood cell (WBC) types are associated with the determination of blood diseases, and therefore accurate classification assures accurate therapy. In this study, we applied deep convolution networks on microscopy hyperspectral images for WBC classification. We proposed a 3d convolutional networks named deep hyper that enables learning spectral and spatial features by itself to make fully use of three dimensional hyperspectral data for white blood cell classification. In order to enhance the representative power of the model in an efficient manner, we integrated a 3d attention module with the last block of the model to put more emphasis on important features. The overall classification accuracy of deep hyper can achieve 96% compared to 90% of the machine learning based method and the attention module allows deep hyper to achieve the best performance of 97.72%. Furthermore, we explored the correlation between spectral features and WBC classification performances to present that hyperspectral characteristics play an important role in classifying specific type of WBC. These findings demonstrate that combing microscopy hyperspectral image with deep convolution networks is beneficial for blood smear analysis especially white blood cell classification.

关键词： White blood cell Hyperspectral image 3d convolutional networks Attention networks

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：