NSTL国家科技图书文献中心

1. Vector Quantized Underwater Image Enhancement With Transformers NSTL国家科技图书文献中心

Xueyan Ding | Yixin Sui ... - 《IEEE Journal of Oceanic Engineering》 - 2025,50(1) - 136～149 - 共14页

摘要： representations of underwater images through a vector quantized | , we propose a vector* quantized underwater image* | vector quantized generative network and an axial flow | -guided latent transformer. The vector quantized | Due to the complexity of underwater imaging

关键词： Vectors | Transformers | Image color analysis | Codes | Feature extraction | Image enhancement | Convolutional neural networks | Complexity theory | Image reconstruction | Generative adversarial networks

2. Quaternion Vector Quantized Variational Autoencoder NSTL国家科技图书文献中心

Hui Luo | Xin Liu ... - 《IEEE signal processing letters》 - 2025,32 - 151～155 - 共5页

摘要：Vector quantized variational autoencoders, as | discrete representations* by quantizing continuous latent* | variants of variational autoencoders, effectively capture | spaces and are widely used in generative tasks. However | , these models still face limitations in handling

关键词： Quaternions | Vectors | Image reconstruction | Convolution | Neural networks | Face recognition | Quantization (signal) | Decoding | Training | Indexes

3. Anomaly detection scheme for lung CT images using vector quantized variational auto-encoder with support vector data description NSTL国家科技图书文献中心

Gao, Zhihui | Nakayama, Ryohei ... - 《Radiological physics and technology》 - 2025,18(1) - 17～26 - 共10页

摘要： introduced to VQ-VAE (Vector* Quantized-Variational Auto* | this study, SVDD (Support Vector* Data Description* | replaced by the embedding representations. The data of | This study aims to develop an anomaly | -detection scheme for lesions in CT images. Our database

关键词： Anomaly detection | Lung CT images | VQ-VAE | SVDD

4. Trajectory Forecasting Through Low-Rank Adaptation of Discrete Latent Codes NSTL国家科技图书文献中心

Riccardo Benaglia | Angelo Porrello ... - 《Pattern Recognition,Part XVI》 - International Conference on Pattern Recognition - 2025, - 236～251 - 共16页

摘要： leveraging Vector* Quantized Variational Autoencoders (VQ* | tailored latent representations* for each example. In a* | Trajectory forecasting is crucial for video | surveillance analytics, as it enables the anticipation of | future movements for a set of agents, e.g., basketball

关键词： Trajectory forecasting | Vector quantization

5. QSIM: A Quantum-inspired hierarchical semantic interaction model for text classification NSTL国家科技图书文献中心

Gao H. | Zhang P.... - 《Neurocomputing》 - 2025,611(Jan.1) - 1.1～1.15 - 共15页

摘要： constructed quantized* representations of language systems in* | the unit vector* in the two-dimensional minimum* | © 2024Semantic interaction modeling is a | fundamental technology in natural language understanding | that guides models to extract deep semantic

关键词： Attention mechanism | Entanglement entropy | Quantum language model | Schmidt decomposition

6. Unsupervised EEG-Based Seizure Anomaly Detection with Denoising Diffusion Probabilistic Models NSTL国家科技图书文献中心

Wang, Jiale | Sun, Mengxue ... - 《International Journal of Neural Systems》 - 2024,34(9) - 共14页

摘要： spectrograms are encoded into vector-quantized | representations, which enables powerful and efficient DDPM while | While many seizure detection methods have | demonstrated great accuracy, their training necessitates a | substantial volume of labeled data. To address this issue

关键词： Seizure detection | denoising diffusion probabilistic models | anomaly detection | unsupervised learning | vector-quantized representations

7. Patch-wise vector quantization for unsupervised medical anomaly detection NSTL国家科技图书文献中心

Kim T. | Jeong I.... - 《Pattern recognition letters》 - 2024,184(Aug.) - 205～211 - 共7页

摘要： vector-quantized representations of normal “patches | pathology detection method called Patch-wise Vector | incorporating vector* similarity metric. Furthermore, we* | collapse” problem of vector* quantization by proposing top* | ? 2024 Elsevier B.V.Radiography images

关键词： Anomaly detection | Computer vision | Medical imaging | Representation learning

8. ε-ViLM : Efficient Video-Language Model via Masked Video Modeling with Semantic Vector-Quantized Tokenizer NSTL国家科技图书文献中心

Jacob Zhiyuan Fang | Skyler Zheng ... - 《2024 IEEE/CVF Winter Conference on Applications of Computer Vision Workshops, [v.1]: WACVW 2024, Waikoloa, Hawaii, USA, 1-6 January 2024》 - IEEE/CVF Conference on Applications of Computer Vision Workshops - 2024, - 529～540 - 共12页

摘要：, produced by the pre-trained vector-quantized* tokenizer* | (MVM) schema, assisted with a semantic vector | -quantized* tokenizer: In particular, our E-ViLM learns to* | representations from Video-Language corpus and generalize well | To build scalable models for the challenging

关键词： Visualization | Computational modeling | Architecture | Semantics | Buildings | Focusing | Question answering (information retrieval)

9. Robust unsupervised-learning based crack detection for stamped metal products NSTL国家科技图书文献中心

Zhang, Penghua | Ryu, Hojun ... - 《Journal of Manufacturing Systems》 - 2024,73 - 65～74 - 共10页

摘要： approach leverages the Vector* Quantized-Variational* | , latent features at different scales are quantized* into* | discrete representations* using a codebook. To learn the* | distribution of these discrete representations* from non-crack* | highest dissimilarity. The edited representations* are*

关键词： Unsupervised learning | Crack detection | Reconstruction | Stamped metal product | VISION

10. Enhancing Image Generation by fusing Auto Encoder & Transformative Generation Approach NSTL国家科技图书文献中心

Eliza Femi Sherley S | Sanjay T ... - 《2024 International Conference on Electrical Electronics and Computing Technologies》 - International Conference on Electrical Electronics and Computing Technologies - 2024, - 1～6 - 共6页

摘要： novel approach by fusing GPT with Vector Quantized | derive discrete latent representations* from input* | images. These latent representations, incorporating | the latent space representations* obtained from VQ* | Generative Pre-trained Transformer (GPT

关键词： Adaptation models | Visualization | Image synthesis | Image color analysis | Computational modeling | Gray-scale | Transformers | Data models | Vectors | Optimization

摘要： representations of underwater images through a vector quantized | , we propose a vector quantized underwater image | vector quantized generative network and an axial flow | -guided latent transformer. The vector quantized | Due to the complexity of underwater imaging

关键词： Vectors | Transformers | Image color analysis | Codes | Feature extraction | Image enhancement | Convolutional neural networks | Complexity theory | Image reconstruction | Generative adversarial networks

摘要：Vector quantized variational autoencoders, as | discrete representations by quantizing continuous latent | variants of variational autoencoders, effectively capture | spaces and are widely used in generative tasks. However | , these models still face limitations in handling

关键词： Quaternions | Vectors | Image reconstruction | Convolution | Neural networks | Face recognition | Quantization (signal) | Decoding | Training | Indexes

摘要： introduced to VQ-VAE (Vector Quantized-Variational Auto | this study, SVDD (Support Vector Data Description | replaced by the embedding representations. The data of | This study aims to develop an anomaly | -detection scheme for lesions in CT images. Our database

关键词： Anomaly detection | Lung CT images | VQ-VAE | SVDD

摘要： leveraging Vector Quantized Variational Autoencoders (VQ | tailored latent representations for each example. In a | Trajectory forecasting is crucial for video | surveillance analytics, as it enables the anticipation of | future movements for a set of agents, e.g., basketball

关键词： Trajectory forecasting | Vector quantization

摘要： constructed quantized representations of language systems in | the unit vector in the two-dimensional minimum | © 2024Semantic interaction modeling is a | fundamental technology in natural language understanding | that guides models to extract deep semantic

关键词： Attention mechanism | Entanglement entropy | Quantum language model | Schmidt decomposition

摘要： spectrograms are encoded into vector-quantized | representations, which enables powerful and efficient DDPM while | While many seizure detection methods have | demonstrated great accuracy, their training necessitates a | substantial volume of labeled data. To address this issue

关键词： Seizure detection | denoising diffusion probabilistic models | anomaly detection | unsupervised learning | vector-quantized representations

摘要： vector-quantized representations of normal “patches | pathology detection method called Patch-wise Vector | incorporating vector similarity metric. Furthermore, we | collapse” problem of vector quantization by proposing top | ? 2024 Elsevier B.V.Radiography images

关键词： Anomaly detection | Computer vision | Medical imaging | Representation learning

摘要：, produced by the pre-trained vector-quantized tokenizer | (MVM) schema, assisted with a semantic vector | -quantized tokenizer: In particular, our E-ViLM learns to | representations from Video-Language corpus and generalize well | To build scalable models for the challenging

关键词： Visualization | Computational modeling | Architecture | Semantics | Buildings | Focusing | Question answering (information retrieval)

摘要： approach leverages the Vector Quantized-Variational | , latent features at different scales are quantized into | discrete representations using a codebook. To learn the | distribution of these discrete representations from non-crack | highest dissimilarity. The edited representations are

关键词： Unsupervised learning | Crack detection | Reconstruction | Stamped metal product | VISION

摘要： novel approach by fusing GPT with Vector Quantized | derive discrete latent representations from input | images. These latent representations, incorporating | the latent space representations obtained from VQ | Generative Pre-trained Transformer (GPT

关键词： Adaptation models | Visualization | Image synthesis | Image color analysis | Computational modeling | Gray-scale | Transformers | Data models | Vectors | Optimization

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： representations of underwater images through a vector quantized | , we propose a vector* quantized underwater image* | vector quantized generative network and an axial flow | -guided latent transformer. The vector quantized | Due to the complexity of underwater imaging

摘要：Vector quantized variational autoencoders, as | discrete representations* by quantizing continuous latent* | variants of variational autoencoders, effectively capture | spaces and are widely used in generative tasks. However | , these models still face limitations in handling

摘要： introduced to VQ-VAE (Vector* Quantized-Variational Auto* | this study, SVDD (Support Vector* Data Description* | replaced by the embedding representations. The data of | This study aims to develop an anomaly | -detection scheme for lesions in CT images. Our database

摘要： leveraging Vector* Quantized Variational Autoencoders (VQ* | tailored latent representations* for each example. In a* | Trajectory forecasting is crucial for video | surveillance analytics, as it enables the anticipation of | future movements for a set of agents, e.g., basketball

摘要： constructed quantized* representations of language systems in* | the unit vector* in the two-dimensional minimum* | © 2024Semantic interaction modeling is a | fundamental technology in natural language understanding | that guides models to extract deep semantic

摘要： vector-quantized representations of normal “patches | pathology detection method called Patch-wise Vector | incorporating vector* similarity metric. Furthermore, we* | collapse” problem of vector* quantization by proposing top* | ? 2024 Elsevier B.V.Radiography images

摘要：, produced by the pre-trained vector-quantized* tokenizer* | (MVM) schema, assisted with a semantic vector | -quantized* tokenizer: In particular, our E-ViLM learns to* | representations from Video-Language corpus and generalize well | To build scalable models for the challenging

摘要： novel approach by fusing GPT with Vector Quantized | derive discrete latent representations* from input* | images. These latent representations, incorporating | the latent space representations* obtained from VQ* | Generative Pre-trained Transformer (GPT