NSTL国家科技图书文献中心

1. ViFi: A Video Finding System at Video Browser Showdown 2025 NSTL国家科技图书文献中心

Khanh-An C. Quan | Qui Ngoc Nguyen ... - 《MultiMedia Modeling,Part V》 - International Conference on MultiMedia Modeling - 2025, - 378～384 - 共7页

摘要： custom keyframe* extraction stage. While neighboring* | extracted keyframe* from the dataset like other previous* | approaches, we found that the given keyframe* might have* | This paper presents ViFi - a Video Finding | System for Video Browser Showdown 2025. Our retrieval

关键词： Video retrieval system | Interactive retrieval system | Multimodal retrieval

2. Adaptive Keyframe Selection Strategy of Visual SLAM in Complex Poses NSTL国家科技图书文献中心

Zu, Linan | Wei, Chengrui ... - 《IEEE sensors journal》 - 2025,25(1) - 1756～1767 - 共12页

摘要： purpose of achieving adaptive compensation in keyframe | Aiming at the accuracy problem of pose | estimation under complex poses of indoor unmanned aerial | vehicles (UAVs), this article proposes a dynamic | threshold setting method based on vision sensor pose

关键词： Cameras | Simultaneous localization and mapping | Sensors | Accuracy | Pose estimation | Visualization | Real-time systems | Uncertainty | Location awareness | Feature extraction...

3. KS-FuseNet: An Efficient Action Recognition Method Based on Keyframe Selection and Feature Fusion NSTL国家科技图书文献中心

Keming Mao | Yilong Xiao ... - 《Pattern Recognition and Computer Vision,Part VII》 - Chinese Conference on Pattern Recognition and Computer Vision - 2025, - 540～553 - 共14页

摘要： an action recognition approach grounded in keyframe | core modules. The keyframe* screening module employs* | action feature extraction* on the differentiated* | Addressing the challenge of effectively | capturing features in contemporary video tasks, we propose

关键词： Action recognition | Spatial-temporal | Feature fusion | Keyframe selection

4. A LiDAR Odometry With Multi-Metric Feature Association and Contribution Constraint Selection NSTL国家科技图书文献中心

Nuo Li | Yiqing Yao ... - 《IEEE robotics and automation letters》 - 2025,10(1) - 756～763 - 共8页

摘要：. Initially, we propose a feature extraction* method centered* | approach for keyframe* registration. This method leverages* | accuracy and speed of keyframe* matching. Additionally* | LiDAR-based simultaneous localization and | mapping (SLAM) is crucial for achieving accurate pose

关键词： Feature extraction | Laser radar | Point cloud compression | Optimization | Simultaneous localization and mapping | Pose estimation | Odometry | Accuracy | Image segmentation | Real-time systems

5. A two-stream sign language recognition network based on keyframe extraction method NSTL国家科技图书文献中心

Tianyu Liu | Tangfei Tao ... - 《Expert Systems with Application》 - 2024,253(Nov.) - 124268.1～124268.17 - 共17页

摘要： of keyframe* extraction methods and two-stream* | key frame extraction* that efficiently summarizes* | Sign language recognition (SLR) tasks are | typically performed on a large number of continuous frames | of sign language videos, making it challenging to

关键词： Attention mechanism | Keyframe extraction | Sign language recognition

6. A method for recognising wrong actions of martial arts athletes based on keyframe extraction NSTL国家科技图书文献中心

Zhiqiang Li - 《International Journal of Biometrics》 - 2024,16(3/4) - 256～271 - 共16页

摘要： on keyframe* extraction. Firstly, the optical flow* | is optimised to obtain the final keyframe* sequence* | In order to improve the accuracy of incorrect | action recognition and shorten the time required for | action recognition, the paper proposes a method for

关键词： keyframe extraction | optical flow method | Lens adaptation | K-means clustering algorithm | Euclidean distance

7. Integrating Keyframe Extraction and MoveNet Thunder for Robust Video Pose Estimation and Classification NSTL国家科技图书文献中心

Samarth Hanchinamani | Pratiksha Tigadi ... - 《2024 5th International Conference for Emerging Technology: 5th International Conference for Emerging Technology (INCET 2024), 24-26 May 2024, Belgaum, India》 - International Conference for Emerging Technology - 2024, - 1～6 - 共6页

摘要： keyframe extraction, convolutional neural networks (CNN | applicability. The fusion of keyframe extraction, CNN, and | This paper introduces a new method to improve | the efficiency of video pose estimation, including | ), and the MoveNet Thunder model. Keyframes, which are

关键词： Adaptation models | Accuracy | Three-dimensional displays | Pose estimation | Computer architecture | Streaming media | Feature extraction

8. OtoFrame: A Software Solution for Keyframe Extraction in Otology Endoscopic Videos NSTL国家科技图书文献中心

Jobanpreet Singh | Mark Prince ... - 《2024 29th International Conference on Automation and Computing》 - International Conference on Automation and Computing - 2024, - 1～6 - 共6页

摘要：. Conventional keyframe* extraction techniques often miss* | employs advanced image extraction* techniques upon* | guarantees precise extraction* of keyframes from otology* | A novel software application to address a | critical industrial challenge is developed to extract

关键词： Computer vision | Accuracy | Transfer learning | Supervised learning | Medical services | Propulsion | Transformers | Software | Standards | Videos

9. MDJ: A multi-scale difference joint keyframe extraction algorithm for infrared surveillance video action recognition NSTL国家科技图书文献中心

Feng, Zhiqiang | Wang, Xiaogang ... - 《Digital Signal Processing》 - 2024,148 - ARTN 104469～ - 共13页 - 被引量：1

摘要： difference joint key frame extraction* algorithm for action* | Many action recognition methods require | significant computational resources to achieve good results | on unedited videos. However, their performance on | infrared videos, which contain less information, is often

关键词： Infrared surveillance video | Information entropy | Multi-scale differences | Action recognition

10. An Approach for Video Summarization based on Unsupervised Learning using Deep Semantic Features and Keyframe Extraction NSTL国家科技图书文献中心

V.S. Gangwani | P. L. Ramteke - 《2024 5th International Conference on Image Processing and Capsule Networks: ICIPCN 2024, Dhulikhel, Nepal, 3-4 July 2024, [v.1]》 - International Conference on Image Processing and Capsule Networks - 2024, - 520～527 - 共8页

摘要： problem with video information extraction* is that* | accurate feature extraction* from video frames. To* | utilization of vision-based feature extraction* yields* | Digital videos have emerged as a prominent and | influential medium for distributing information, with its

摘要： custom keyframe* extraction stage. While neighboring* | extracted keyframe* from the dataset like other previous* | approaches, we found that the given keyframe* might have* | This paper presents ViFi - a Video Finding | System for Video Browser Showdown 2025. Our retrieval

关键词： Video retrieval system | Interactive retrieval system | Multimodal retrieval

摘要： purpose of achieving adaptive compensation in keyframe | Aiming at the accuracy problem of pose | estimation under complex poses of indoor unmanned aerial | vehicles (UAVs), this article proposes a dynamic | threshold setting method based on vision sensor pose

关键词： Cameras | Simultaneous localization and mapping | Sensors | Accuracy | Pose estimation | Visualization | Real-time systems | Uncertainty | Location awareness | Feature extraction...

摘要： an action recognition approach grounded in keyframe | core modules. The keyframe* screening module employs* | action feature extraction* on the differentiated* | Addressing the challenge of effectively | capturing features in contemporary video tasks, we propose

关键词： Action recognition | Spatial-temporal | Feature fusion | Keyframe selection

摘要：. Initially, we propose a feature extraction* method centered* | approach for keyframe* registration. This method leverages* | accuracy and speed of keyframe* matching. Additionally* | LiDAR-based simultaneous localization and | mapping (SLAM) is crucial for achieving accurate pose

关键词： Feature extraction | Laser radar | Point cloud compression | Optimization | Simultaneous localization and mapping | Pose estimation | Odometry | Accuracy | Image segmentation | Real-time systems

摘要： of keyframe* extraction methods and two-stream* | key frame extraction* that efficiently summarizes* | Sign language recognition (SLR) tasks are | typically performed on a large number of continuous frames | of sign language videos, making it challenging to

关键词： Attention mechanism | Keyframe extraction | Sign language recognition

摘要： on keyframe* extraction. Firstly, the optical flow* | is optimised to obtain the final keyframe* sequence* | In order to improve the accuracy of incorrect | action recognition and shorten the time required for | action recognition, the paper proposes a method for

关键词： keyframe extraction | optical flow method | Lens adaptation | K-means clustering algorithm | Euclidean distance

摘要： keyframe extraction, convolutional neural networks (CNN | applicability. The fusion of keyframe extraction, CNN, and | This paper introduces a new method to improve | the efficiency of video pose estimation, including | ), and the MoveNet Thunder model. Keyframes, which are

关键词： Adaptation models | Accuracy | Three-dimensional displays | Pose estimation | Computer architecture | Streaming media | Feature extraction

摘要：. Conventional keyframe* extraction techniques often miss* | employs advanced image extraction* techniques upon* | guarantees precise extraction* of keyframes from otology* | A novel software application to address a | critical industrial challenge is developed to extract

关键词： Computer vision | Accuracy | Transfer learning | Supervised learning | Medical services | Propulsion | Transformers | Software | Standards | Videos

摘要： difference joint key frame extraction* algorithm for action* | Many action recognition methods require | significant computational resources to achieve good results | on unedited videos. However, their performance on | infrared videos, which contain less information, is often

关键词： Infrared surveillance video | Information entropy | Multi-scale differences | Action recognition

摘要： problem with video information extraction* is that* | accurate feature extraction* from video frames. To* | utilization of vision-based feature extraction* yields* | Digital videos have emerged as a prominent and | influential medium for distributing information, with its

关键词： Representation learning | Image segmentation | Semantics | Object detection | Feature extraction | Motion pictures | Information retrieval

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： custom keyframe extraction stage. While neighboring | extracted keyframe from the dataset like other previous | approaches, we found that the given keyframe might have | This paper presents ViFi - a Video Finding | System for Video Browser Showdown 2025. Our retrieval

关键词： Video retrieval system | Interactive retrieval system | Multimodal retrieval

摘要： purpose of achieving adaptive compensation in keyframe | Aiming at the accuracy problem of pose | estimation under complex poses of indoor unmanned aerial | vehicles (UAVs), this article proposes a dynamic | threshold setting method based on vision sensor pose

关键词： Cameras | Simultaneous localization and mapping | Sensors | Accuracy | Pose estimation | Visualization | Real-time systems | Uncertainty | Location awareness | Feature extraction...

摘要： an action recognition approach grounded in keyframe | core modules. The keyframe screening module employs | action feature extraction on the differentiated | Addressing the challenge of effectively | capturing features in contemporary video tasks, we propose

关键词： Action recognition | Spatial-temporal | Feature fusion | Keyframe selection

摘要：. Initially, we propose a feature extraction method centered | approach for keyframe registration. This method leverages | accuracy and speed of keyframe matching. Additionally | LiDAR-based simultaneous localization and | mapping (SLAM) is crucial for achieving accurate pose

关键词： Feature extraction | Laser radar | Point cloud compression | Optimization | Simultaneous localization and mapping | Pose estimation | Odometry | Accuracy | Image segmentation | Real-time systems

摘要： of keyframe extraction methods and two-stream | key frame extraction that efficiently summarizes | Sign language recognition (SLR) tasks are | typically performed on a large number of continuous frames | of sign language videos, making it challenging to

关键词： Attention mechanism | Keyframe extraction | Sign language recognition

摘要： on keyframe extraction. Firstly, the optical flow | is optimised to obtain the final keyframe sequence | In order to improve the accuracy of incorrect | action recognition and shorten the time required for | action recognition, the paper proposes a method for

关键词： keyframe extraction | optical flow method | Lens adaptation | K-means clustering algorithm | Euclidean distance

摘要： keyframe extraction, convolutional neural networks (CNN | applicability. The fusion of keyframe extraction, CNN, and | This paper introduces a new method to improve | the efficiency of video pose estimation, including | ), and the MoveNet Thunder model. Keyframes, which are

关键词： Adaptation models | Accuracy | Three-dimensional displays | Pose estimation | Computer architecture | Streaming media | Feature extraction

摘要：. Conventional keyframe extraction techniques often miss | employs advanced image extraction techniques upon | guarantees precise extraction of keyframes from otology | A novel software application to address a | critical industrial challenge is developed to extract

关键词： Computer vision | Accuracy | Transfer learning | Supervised learning | Medical services | Propulsion | Transformers | Software | Standards | Videos

摘要： difference joint key frame extraction algorithm for action | Many action recognition methods require | significant computational resources to achieve good results | on unedited videos. However, their performance on | infrared videos, which contain less information, is often

关键词： Infrared surveillance video | Information entropy | Multi-scale differences | Action recognition

摘要： problem with video information extraction is that | accurate feature extraction from video frames. To | utilization of vision-based feature extraction yields | Digital videos have emerged as a prominent and | influential medium for distributing information, with its

关键词： Representation learning | Image segmentation | Semantics | Object detection | Feature extraction | Motion pictures | Information retrieval

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： custom keyframe* extraction stage. While neighboring* | extracted keyframe* from the dataset like other previous* | approaches, we found that the given keyframe* might have* | This paper presents ViFi - a Video Finding | System for Video Browser Showdown 2025. Our retrieval

摘要： an action recognition approach grounded in keyframe | core modules. The keyframe* screening module employs* | action feature extraction* on the differentiated* | Addressing the challenge of effectively | capturing features in contemporary video tasks, we propose

摘要：. Initially, we propose a feature extraction* method centered* | approach for keyframe* registration. This method leverages* | accuracy and speed of keyframe* matching. Additionally* | LiDAR-based simultaneous localization and | mapping (SLAM) is crucial for achieving accurate pose

摘要： of keyframe* extraction methods and two-stream* | key frame extraction* that efficiently summarizes* | Sign language recognition (SLR) tasks are | typically performed on a large number of continuous frames | of sign language videos, making it challenging to

摘要： on keyframe* extraction. Firstly, the optical flow* | is optimised to obtain the final keyframe* sequence* | In order to improve the accuracy of incorrect | action recognition and shorten the time required for | action recognition, the paper proposes a method for

摘要：. Conventional keyframe* extraction techniques often miss* | employs advanced image extraction* techniques upon* | guarantees precise extraction* of keyframes from otology* | A novel software application to address a | critical industrial challenge is developed to extract

摘要： difference joint key frame extraction* algorithm for action* | Many action recognition methods require | significant computational resources to achieve good results | on unedited videos. However, their performance on | infrared videos, which contain less information, is often

摘要： problem with video information extraction* is that* | accurate feature extraction* from video frames. To* | utilization of vision-based feature extraction* yields* | Digital videos have emerged as a prominent and | influential medium for distributing information, with its