NSTL国家科技图书文献中心

1. CNN Mixture-of-Depths NSTL国家科技图书文献中心

Rinor Cakaj | Jens Mehnert ... - 《Computer Vision - ACCV 2024,Part VII》 - Asian Conference on Computer Vision - 2025, - 148～166 - 共19页

摘要： that require dynamic computation graphs, CNN* MoD uses* | -tuning. CNN* MoD either matches the performance of* | We introduce Mixture-of-Depths (MoD) for | Convolutional Neural Networks (CNNs), a novel approach that | enhances the computational efficiency of CNNs by

关键词： CNN | Mixture-of-Depths | Computational efficiency | Inference speed

2. Lightweight Single Image Super-Resolution Network Integrating CNN and Transformer NSTL国家科技图书文献中心

Kai Zhu | Li Chen - 《Pattern Recognition,Part XXXII》 - International Conference on Pattern Recognition - 2025, - 212～225 - 共14页

摘要： based on CNN* and Transformer architectures. However* | an Alternating CNN* Transformer Block and an* | Integrative CNN* Efficient Transformer for single image super* | combining CNN* within and between Transformer modules. In* | Gated CNN* and Parallel Channel Attention, aiming to*

关键词： SISR | CNN | Transformer | Attention | Deep learning

3. Rolling bearing fault diagnosis based on VMD-CNN-BiLSTM NSTL国家科技图书文献中心

Zhiqiang Wu - 《Fourth International Conference on Computer Vision,Application,and Algorithm (CVAA 2024)》 - International Conference on Computer Vision,Application,and Algorithm - 2025, - 1348622.1～1348622.7 - 共7页

摘要：), Convolutional Neural Network (CNN), and Bidirectional Long | -CNN-BiLSTM, for rolling bearing fault diagnosis. VMD | endpoint effects. The CNN* layer extracts local features* | innovation lies in integrating VMD with CNN* and BiLSTM* | demonstrate that the VMD-CNN-BiLSTM model achieves a

关键词： Rolling bearings | Fault diagnosis | VMD | CNN | BiLSTM

4. PolSAR Image Classification Using Superpixel Profile and CNN NSTL国家科技图书文献中心

Nabajyoti Das | Swarnajyoti Patra ... - 《Pattern Recognition,Part II》 - International Conference on Pattern Recognition - 2025, - 320～334 - 共15页

摘要： (CNN) based frameworks are being applied to | performance of CNN* for PolSAR image classification is* | features. Although CNN* automatically extracts abstract* | into a CNN* model for classification. The experiment* | comparison to the state-of-the-art CNN* model.*

关键词： PolSAR images | Superpixel profile | Convolutional neural network (CNN)

5. FARSE-CNN: Fully Asynchronous, Recurrent and Sparse Event-Based CNN NSTL国家科技图书文献中心

Riccardo Santambrogi... | Marco Cannici ... - 《Computer Vision - ECCV 2024,Part LIV》 - European Conference on Computer Vision - 2025, - 1～18 - 共18页

摘要： Asynchronous, Recurrent and Sparse Event-Based CNN (FARSE-CNN | recognition. FARSE-CNN* achieves similar or better* | released at https://github.com/AIRLab-POLIMI/farse-cnn. | Event cameras are neuromorphic image sensors | that respond to per-pixel brightness changes

关键词： Event cameras | Deep learning architectures

6. Application of CNN in condition monitoring of high voltage switchgear NSTL国家科技图书文献中心

Lei Sun | Yao Li ... - 《International Conference on Physics,Photonics,and Optical Engineering (ICPPOE 2024),Part One of Two Parts》 - International Conference on Physics,Photonics,and Optical Engineering - 2025, - 135521B.1～135521B.7 - 共7页

摘要： of Convolutional Neural Networks (CNN) in high | distribution monitoring. Specifically tailored CNN | distribution monitoring, presenting CNN* as a promising* | involve continuous refinement of CNN* architectures and* | Study explores the transformative application

关键词： Convolutional neural networks (CNN) | High-voltage switchgear | State monitoring | Power distribution | Precision | Efficiency | Fault detection

7. Diagnosing ADHD with 1D-CNN: Efficient Analysis of fMRI Data NSTL国家科技图书文献中心

Qurat Ul Ain | Soyiba Jawed - 《Seventeenth International Conference on Machine Vision (ICMV 2024)》 - International Conference on Machine Vision - 2025, - 135170E.1～135170E.8 - 共8页

摘要： Convolution Neural Network (1D-CNN) for the diagnosis of | Independent Component Analysis and then passed to 1D-CNN* for* | Attention Deficit Hyperactivity Disorder (ADHD | ) is a type of neurodevelopmental disease affecting | the mental health of children and adults. So

关键词： ADHD | Deep learning | ICA | 1D-CNN | Rs-fMRI

8. Towards a Lightweight CNN for Semantic Food Segmentation NSTL国家科技图书文献中心

Bastian Munoz | Beatriz Remeseiro ... - 《Progress in Pattern Recognition,Image Analysis,Computer Vision,and Applications,Part I》 - Iberoamerican Congress on Pattern Recognition - 2025, - 1～15 - 共15页

摘要： lightweight CNN* EfficientNet-B1 and the Atrous Spatial* | Semantic food segmentation is an important | task for the development of nutritional systems that | effectively manage daily diets. Recent advances in semantic | segmentation have brought great performance improvements

关键词： Deep network | Lightweights CNN | DeepLabv3+ | Semantic food segmentation

9. Block Cipher Algorithm Identification Based on CNN-Transformer Fusion Model NSTL国家科技图书文献中心

Rongna Xie | Xiaoyu Chen ... - 《Pattern Recognition and Computer Vision,Part XI》 - Chinese Conference on Pattern Recognition and Computer Vision - 2025, - 97～110 - 共14页

摘要： features. The model is based on CNN-Transformer, which | Cryptanalysis is predicated on the recognition | of cipher algorithms, but in practice, researcher | often do not know the cipher algorithm used. This | paper focuses on block cipher algorithms

关键词： Block cipher algorithms identification | Interpretable model | CNN-transformer fusion

10. On the Influence of CNN-Based Feature Learning Modules in Neural Speaker Verification Framework NSTL国家科技图书文献中心

Jahangir Alam | Md Shahidul Alam - 《Speech and Computer,Part II》 - International Conference on Speech and Computer - 2025, - 157～170 - 共14页

摘要：). In this paper, we explore the influence of CNN | neural speaker embedding framework consisting of CNN | network in a cascade arrangement. The CNN-based feature | extraction modules considered are: (i) 1-D CNN, (ii | ) vanilla 2D-CNN, and (iii) Selective Kernel Attention

摘要： that require dynamic computation graphs, CNN* MoD uses* | -tuning. CNN* MoD either matches the performance of* | We introduce Mixture-of-Depths (MoD) for | Convolutional Neural Networks (CNNs), a novel approach that | enhances the computational efficiency of CNNs by

关键词： CNN | Mixture-of-Depths | Computational efficiency | Inference speed

摘要： based on CNN* and Transformer architectures. However* | an Alternating CNN* Transformer Block and an* | Integrative CNN* Efficient Transformer for single image super* | combining CNN* within and between Transformer modules. In* | Gated CNN* and Parallel Channel Attention, aiming to*

关键词： SISR | CNN | Transformer | Attention | Deep learning

摘要：), Convolutional Neural Network (CNN), and Bidirectional Long | -CNN-BiLSTM, for rolling bearing fault diagnosis. VMD | endpoint effects. The CNN* layer extracts local features* | innovation lies in integrating VMD with CNN* and BiLSTM* | demonstrate that the VMD-CNN-BiLSTM model achieves a

关键词： Rolling bearings | Fault diagnosis | VMD | CNN | BiLSTM

摘要： (CNN) based frameworks are being applied to | performance of CNN* for PolSAR image classification is* | features. Although CNN* automatically extracts abstract* | into a CNN* model for classification. The experiment* | comparison to the state-of-the-art CNN* model.*

关键词： PolSAR images | Superpixel profile | Convolutional neural network (CNN)

摘要： Asynchronous, Recurrent and Sparse Event-Based CNN (FARSE-CNN | recognition. FARSE-CNN* achieves similar or better* | released at https://github.com/AIRLab-POLIMI/farse-cnn. | Event cameras are neuromorphic image sensors | that respond to per-pixel brightness changes

关键词： Event cameras | Deep learning architectures

摘要： of Convolutional Neural Networks (CNN) in high | distribution monitoring. Specifically tailored CNN | distribution monitoring, presenting CNN* as a promising* | involve continuous refinement of CNN* architectures and* | Study explores the transformative application

关键词： Convolutional neural networks (CNN) | High-voltage switchgear | State monitoring | Power distribution | Precision | Efficiency | Fault detection

摘要： Convolution Neural Network (1D-CNN) for the diagnosis of | Independent Component Analysis and then passed to 1D-CNN* for* | Attention Deficit Hyperactivity Disorder (ADHD | ) is a type of neurodevelopmental disease affecting | the mental health of children and adults. So

关键词： ADHD | Deep learning | ICA | 1D-CNN | Rs-fMRI

摘要： lightweight CNN* EfficientNet-B1 and the Atrous Spatial* | Semantic food segmentation is an important | task for the development of nutritional systems that | effectively manage daily diets. Recent advances in semantic | segmentation have brought great performance improvements

关键词： Deep network | Lightweights CNN | DeepLabv3+ | Semantic food segmentation

摘要： features. The model is based on CNN-Transformer, which | Cryptanalysis is predicated on the recognition | of cipher algorithms, but in practice, researcher | often do not know the cipher algorithm used. This | paper focuses on block cipher algorithms

关键词： Block cipher algorithms identification | Interpretable model | CNN-transformer fusion

摘要：). In this paper, we explore the influence of CNN | neural speaker embedding framework consisting of CNN | network in a cascade arrangement. The CNN-based feature | extraction modules considered are: (i) 1-D CNN, (ii | ) vanilla 2D-CNN, and (iii) Selective Kernel Attention

关键词： Speaker verification | CNN | Hybrid network | TDNN-LSTM | Cosine scoring | VoxCeleb

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： that require dynamic computation graphs, CNN MoD uses | -tuning. CNN MoD either matches the performance of | We introduce Mixture-of-Depths (MoD) for | Convolutional Neural Networks (CNNs), a novel approach that | enhances the computational efficiency of CNNs by

关键词： CNN | Mixture-of-Depths | Computational efficiency | Inference speed

摘要： based on CNN and Transformer architectures. However | an Alternating CNN Transformer Block and an | Integrative CNN Efficient Transformer for single image super | combining CNN within and between Transformer modules. In | Gated CNN and Parallel Channel Attention, aiming to

关键词： SISR | CNN | Transformer | Attention | Deep learning

摘要：), Convolutional Neural Network (CNN), and Bidirectional Long | -CNN-BiLSTM, for rolling bearing fault diagnosis. VMD | endpoint effects. The CNN layer extracts local features | innovation lies in integrating VMD with CNN and BiLSTM | demonstrate that the VMD-CNN-BiLSTM model achieves a

关键词： Rolling bearings | Fault diagnosis | VMD | CNN | BiLSTM

摘要： (CNN) based frameworks are being applied to | performance of CNN for PolSAR image classification is | features. Although CNN automatically extracts abstract | into a CNN model for classification. The experiment | comparison to the state-of-the-art CNN model.

关键词： PolSAR images | Superpixel profile | Convolutional neural network (CNN)

摘要： Asynchronous, Recurrent and Sparse Event-Based CNN (FARSE-CNN | recognition. FARSE-CNN achieves similar or better | released at https://github.com/AIRLab-POLIMI/farse-cnn. | Event cameras are neuromorphic image sensors | that respond to per-pixel brightness changes

关键词： Event cameras | Deep learning architectures

摘要： of Convolutional Neural Networks (CNN) in high | distribution monitoring. Specifically tailored CNN | distribution monitoring, presenting CNN as a promising | involve continuous refinement of CNN architectures and | Study explores the transformative application

关键词： Convolutional neural networks (CNN) | High-voltage switchgear | State monitoring | Power distribution | Precision | Efficiency | Fault detection

摘要： Convolution Neural Network (1D-CNN) for the diagnosis of | Independent Component Analysis and then passed to 1D-CNN for | Attention Deficit Hyperactivity Disorder (ADHD | ) is a type of neurodevelopmental disease affecting | the mental health of children and adults. So

关键词： ADHD | Deep learning | ICA | 1D-CNN | Rs-fMRI

摘要： lightweight CNN EfficientNet-B1 and the Atrous Spatial | Semantic food segmentation is an important | task for the development of nutritional systems that | effectively manage daily diets. Recent advances in semantic | segmentation have brought great performance improvements

关键词： Deep network | Lightweights CNN | DeepLabv3+ | Semantic food segmentation

摘要： features. The model is based on CNN-Transformer, which | Cryptanalysis is predicated on the recognition | of cipher algorithms, but in practice, researcher | often do not know the cipher algorithm used. This | paper focuses on block cipher algorithms

关键词： Block cipher algorithms identification | Interpretable model | CNN-transformer fusion

摘要：). In this paper, we explore the influence of CNN | neural speaker embedding framework consisting of CNN | network in a cascade arrangement. The CNN-based feature | extraction modules considered are: (i) 1-D CNN, (ii | ) vanilla 2D-CNN, and (iii) Selective Kernel Attention

关键词： Speaker verification | CNN | Hybrid network | TDNN-LSTM | Cosine scoring | VoxCeleb

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： that require dynamic computation graphs, CNN* MoD uses* | -tuning. CNN* MoD either matches the performance of* | We introduce Mixture-of-Depths (MoD) for | Convolutional Neural Networks (CNNs), a novel approach that | enhances the computational efficiency of CNNs by

摘要： based on CNN* and Transformer architectures. However* | an Alternating CNN* Transformer Block and an* | Integrative CNN* Efficient Transformer for single image super* | combining CNN* within and between Transformer modules. In* | Gated CNN* and Parallel Channel Attention, aiming to*

摘要： (CNN) based frameworks are being applied to | performance of CNN* for PolSAR image classification is* | features. Although CNN* automatically extracts abstract* | into a CNN* model for classification. The experiment* | comparison to the state-of-the-art CNN* model.*

摘要： Asynchronous, Recurrent and Sparse Event-Based CNN (FARSE-CNN | recognition. FARSE-CNN* achieves similar or better* | released at https://github.com/AIRLab-POLIMI/farse-cnn. | Event cameras are neuromorphic image sensors | that respond to per-pixel brightness changes

摘要： of Convolutional Neural Networks (CNN) in high | distribution monitoring. Specifically tailored CNN | distribution monitoring, presenting CNN* as a promising* | involve continuous refinement of CNN* architectures and* | Study explores the transformative application

摘要： Convolution Neural Network (1D-CNN) for the diagnosis of | Independent Component Analysis and then passed to 1D-CNN* for* | Attention Deficit Hyperactivity Disorder (ADHD | ) is a type of neurodevelopmental disease affecting | the mental health of children and adults. So

摘要： lightweight CNN* EfficientNet-B1 and the Atrous Spatial* | Semantic food segmentation is an important | task for the development of nutritional systems that | effectively manage daily diets. Recent advances in semantic | segmentation have brought great performance improvements