检索结果-内蒙古大学图书馆

SEDNet: Real-Time Semantic Segmentation Algorithm Based on STDC

INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS 2025年第1期2025卷

作者： Ma, Sugang Zhao, Ziyi Hou, Zhiqiang Yu, Wangsheng Yang, Xiaobao Zhao, Xiangmo Xian Univ Posts & Telecommun Sch Comp Sci & Technol Xian Peoples R China Changan Univ Sch Informat Engn Xian Peoples R China Air Force Engn Univ Sch Informat & Nav Xian Peoples R China

Recently, deep convolutional neural networks (DCNN) have been widely used in semantic segmentation tasks and have achieved high segmentation accuracy. However, most algorithms based on DCNN have high computational complexity, making them unsuitable for real-time segmentation. To solve this problem, this paper proposes a real-time semantic segmentation algorithm based on the STDC network. The algorithm adopts an "encoder-decoder" embedded in a U-shaped architecture to realize real-time segmentation while maintaining high accuracy. Following the encoder, a mixed pooling attention module is designed to expand the receptive field, enhancing the network model's learning ability in complex scenarios. Then, a feature fusion module is used for combining features from different stages, and channel attention based on atrous convolution is employed to expand the receptive field and avoid dimensionality reduction learning. Finally, a Tversky-based detail loss function is used to encode more spatial details. The proposed algorithm was extensively tested on the challenging Cityscapes and CamVid datasets, and the experimental results showed that the proposed algorithm obtained 76.4% and 72.8% of mIoU, respectively. Meanwhile, our algorithm achieves 105.2 FPS and 165.6 FPS inference speed with a single NVIDIA GTX 1080Ti GPU, meeting the real-time segmentation requirements. The proposed algorithm can conduct real-time segmentation while maintaining high accuracy, achieving a good balance between accuracy and speed.

关键词： deep convolutional neural network encoder-decoder real-time semantic segmentation STDC

来源：评论

学校读者我要写书评

暂无评论

Adaptively forget with crossmodal and textual distillation for class-incremental video captioning

引用

NEUROCOMPUTING 2025年 624卷

作者： Xiong, Huiyu Wang, Lanxiao Qiu, Heqian Zhao, Taijin Qiu, Benliu Li, Hongliang Univ Elect Sci & Technol China Sch Informat & Commun Engn Chengdu 611731 Peoples R China

With the increasing volume of video data, it is critical that intelligent systems have the ability to continuously analyze the increasing number of videos while avoiding the usage of data from early tasks. These video data are difficult to store, and training models from scratch face privacy and high energy consumption pitfalls. However, existing research has primarily focused on incremental learning for image and video classification, with limited attention given to analyzing multimodal inputs and leveraging the spatio-temporal contextual information embedded in videos to generate meaningful captions. In this paper, we first present the C lass-Incremental V ideo C aptioning (CI-VC) task. There are two difficulties in this task, which are the stability-plasticity dilemma faced in processing textual information with spatio-temporal states and correlating video-text inter-modal information, respectively. To fill this unexplored area, we primitively propose a method to A daptively F orget with crossmodal and textual D istillation for class-incremental video captioning (AFD). Considering the unique encoder-decoder networks of captioning and the privacy of input data, we refer to the three stages without playback of old samples: visual encoding, textual decoding and network supervision. In the phase of learning new knowledge, we use F ine- g rained S ensitivity S election (FgSS) to refine the selection of stored knowledge from the old model, forgetting some irrelevant information to make room for the memory of new classes. In order to memorize the text information during decoding, we design G rowing Word Embedding (c) with elastic capacity to store the embedding tokens of text that have already been seen as textual meta-signal pairs to be retrieved the lexical information during further caption generation. Meanwhile, the D ual-level K nowledge D istillation (DlKD) is performed at both cross-modal semantic information and the textual final output to constrain from the per

关键词： Class-incremental learning Video captioning encoder-decoder Multimodal attention Knowledge distillation

来源：评论

学校读者我要写书评

暂无评论

Recurrent fusion transformer for image captioning

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2025年第1期19卷 1-9页

作者： Mou, Zhenping Yuan, Qiao Song, Tianqi Chongqing Normal Univ Sch Comp & Informat Sci Chongqing 401331 Peoples R China Chongqing Univ Sci & Technol Coll Intelligent Technol & Engn Chongqing 401331 Peoples R China

Image captioning describes the visual content of a given image by using natural language sentences. However, in the existing image captioning model, the encoder only describes the image content from a specific pattern, and cannot fully understand the semantic sequence information of the input image. In this paper, we propose a multimodal recurrent fusion block (RF-Block), which uses a new recurrent attention and combines gated recurrent networks to capture feature correlation information. Also, a stack multiple feature fusion blocks is used can better enhance the relationship between higher-level features. Finally, the feature fusion block is inserted into the Transformer to form a recurrent fusion transformer (RCT), which can improve the performance of the image captioning model. Experimental results show that the proposed model is better than the traditional encoder-decoder image captioning model and provides comparable performance to the most advanced models.

关键词： Image captioning Recurrent fusion block Transformer encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

CFPNet: Complementary Feature Perception Network for Underwater Image Enhancement

引用

IEEE JOURNAL OF OCEANIC ENGINEERING 2025年第1期50卷 150-163页

作者： Fu, Xianping Qin, Wenqiang Li, Fengqi Xu, Fengqiang Yan, Xiaohong Dalian Maritime Univ Sch Informat Sci & Technol Dalian 116026 Peoples R China Dalian Jiaotong Univ Sch Software Dalian 116600 Peoples R China Dalian Key Lab Blockchain Technol & Applicat Dalian 116028 Peoples R China

Images shot underwater are usually characterized by global nonuniform information loss due to selective light absorption and scattering, resulting in various degradation problems, such as color distortion and low visibility. Recently, deep learning has drawn much attention in the field of underwater image enhancement (UIE) for its powerful performance. However, most deep learning-based UIE models rely on either pure convolutional neural network (CNN) or pure transformer, which makes it challenging to enhance images while maintaining local representations and global features simultaneously. In this article, we propose a novel complementary feature perception network (CFPNet), which embeds the transformer into the classical CNN-based UNet3+. The core idea is to fuse the advantages of CNN and transformer to obtain satisfactory high-quality underwater images that can naturally perceive local and global features. CFPNet employs a novel dual encoder structure of the CNN and transformer in parallel, while the decoder is composed of one trunk decoder and two auxiliary decoders. First, we propose the regionalized two-stage vision transformer that can progressively eliminate the variable levels of degradation in a coarse-to-fine manner. Second, we design the full-scale feature fusion module to explore sufficient information by merging the multiscale features. In addition, we propose an auxiliary feature guided learning strategy that utilizes reflectance and shading maps to guide the generation of the final results. The advantage of this strategy is to avoid repetitive and ineffective learning of the model, and to accomplish color correction and deblurring tasks more efficiently. Experiments demonstrate that our CFPNet can obtain high-quality underwater images and show superior performance compared to the state-of-the-art UIE methods qualitatively and quantitatively.

关键词： Deep learning encoder-decoder image enhancement transformer underwater image

来源：评论

学校读者我要写书评

暂无评论

Deep Feature Blend Attention: A New Frontier in Super Resolution Image Generation

引用

NEUROCOMPUTING 2025年 618卷

作者： Dhanusha, P. B. Muthukumar, A. Lakshmi, A. Kalasalingam Acad Res & Educ Dept Elect & Commun Engn Virudunagar Tamil Nadu India Ramco Inst Technol Dept Elect & Commun Engn Rajapalayam Tamil Nadu India

Super Resolution (SR) images contain more useful information than Low Resolution (LR) images. Generally, SR images are highly preferred over LR images in the medical field due to their high quality. Generally, SR images are affected by many factors, like blur, noise, and decimation. Therefore, the present study proposes Deep Feature Blend Attention Mechanism for generating SR images with optimal outcomes. In this study, a Deep Learning (DL) based encoder-decoder is used to extract the detailed information from the LR image, assist in noise removal, and improve image quality. Using Gaussian blur, an LR image is produced, which is given as input to the encoder-decoder. The Attention mechanism with feature blending is performed to produce the reconstructed SR image. The low and high frequency component issues from feature map are improved using the attention mechanism. Also, feature blend attention mechanism is implied to reconstruct SR images from LR images. The present study utilizes the feature blend technique, which helps select the optimal features and blending to attain the optimal feature set. The study is performed with and without the feature blend technique in attention mechanism to expose the impact of the feature blend technique, and it is used to avoid overfitting issues and produce optimal results. The performance of the proposed system is assessed by using PSNR and SSIM to show the system's efficiency. The proposed system is compared with other state-of-the-art studies to expose the efficacy of the present system.

关键词： Super resolution image generation encoder-decoder Feature blend Attention mechanism

来源：评论

学校读者我要写书评

暂无评论

A deep learning model based on multi-attention mechanism and gated recurrent unit network for photovoltaic power forecasting

引用

COMPUTERS & ELECTRICAL ENGINEERING 2025年 123卷

作者： Yang, Kuo Cai, Yanjie Cheng, Jinrun Shanghai Dianji Univ Sch Mech Engn Shanghai 200444 Peoples R China

Solar energy plays a crucial role in the power grid due to its clean, stable, and cost-effective nature, as well as its significant storage potential. Accurate short-term photovoltaic (PV) power forecasting is essential for effective grid management and dispatching decisions. This study introduces a hybrid deep learning model integrating multiple attention mechanisms and gated recurrent unit networks to forecast PV output power one day in advance. To address the impact of random weather variations and historical PV power data on forecasting accuracy, the model incorporates an input attention mechanism to process input features. Additionally, temporal and spatial attention mechanisms are embedded within the encoder-decoder framework to enhance prediction performance. These mechanisms effectively capture the relationships between historical PV power output and meteorological variables while identifying crucial time-dependent hidden states. The proposed model is validated on a real-world PV dataset, achieving a mean absolute error of 0.0903 under favorable weather conditions, demonstrating a 22.5 % improvement over traditional forecasting methods across various weather classifications. Comparative analyses with other state-of-the-art models confirm that the proposed approach offers superior predictive accuracy.

关键词： Multi-attention mechanism Gated recurrent unit encoder-decoder Photovoltaic power prediction

来源：评论

学校读者我要写书评

暂无评论

Transformer-Based Multi-Target Object Detection and Tracking Framework for Robust Spatio-Temporal Memory in Dynamic Environments

引用

IEEE ACCESS 2025年 13卷 47146-47164页

作者： Alzubi, Tareq Mahmod Mukhtar, Umar Raza Al Balqa Appl Univ Prince Abdullah Bin Ghazi Fac Informat & Commun Te Dept Comp Sci As Salt 19385 Jordan Univ Management & Technol Dept Comp Sci Lahore 54770 Pakistan

Multiple object tracking (MOT) is a critical task in computer vision, applied in areas such as autonomous driving, sports analytics, and human activity recognition. Traditional online MOT approaches, which rely on separate detection and re-identification (re-ID) stages, often suffer from high computational costs and suboptimal accuracy. To address these challenges, we propose MOTTNet, a unified framework that integrates detection and tracklet mapping using transformer-based architectures. MOTTNet leverages a spatiotemporal memory network to store past observations of tracked objects, allowing for more robust identity preservation and trajectory prediction over time. The Target Proposal Module generates object candidates using a transformer encoder-decoder structure, and the memory decoder integrates both proposal and track vectors to predict object locations and identities in real-time, minimizing post-processing. The framework incorporates a multi-scale attention pyramid to handle variations in object scale and utilizes a deformable transformer for generating accurate object proposals. A memory encoding-decoding aggregates object features and associates them across frames, allowing robust detection and identity preservation over time. MOTTNet achieves IDF1 scores of 79.8 and 80.4, MOTA scores of 79.3 and 80.1, and HOTA scores of 73.2 and 72.7, respectively, on on the MOT17 and MOT20 datasets. Additionally, MOTTNet exhibits lower identity switch rates, significantly outperforming existing transformer-based and traditional tracking models.

关键词： Feature extraction Transformers Target tracking Vectors Accuracy Trajectory Object detection Visualization Sports Real-time systems Multi-object tracking memory network transformer multi-scale pyramid encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

FS-Net: Full scale network and adaptive threshold for improving extraction of micro-retinal vessel structures

引用

PATTERN RECOGNITION LETTERS 2025年 189卷 188-194页

作者： Getahun, Melaku N. Rogov, Oleg Y. Dylov, Dmitry V. Somov, Andrey Bouridane, Ahmed Hamoudi, Rifat Skolkovo Inst Sci & Technol Skoltech Moscow Russia Univ Sharjah Ctr Data Analyt & Cybersecur Sharjah U Arab Emirates Univ Sharjah Res Inst Med & Hlth Sci Coll Med Sharjah U Arab Emirates UCL Fac Med Sci Div Surg & Intervent Sci London England Univ Sharjah Biomed Informed Artificial Intelligence Lab BIMAI Lab Sharjah U Arab Emirates Bian Que AI Hong Kong Peoples R China

Retinal vascular segmentation, a widely researched topic in biomedical image processing, aims to reduce the workload of ophthalmologists in treating and detecting retinal disorders. Segmenting retinal vessels presents unique challenges;previous techniques often failed to effectively segment branches and microvascular structures. Recent neural network approaches struggle to balance local and global properties and frequently miss tiny end vessels, hindering the achievement of desired results. To address these issues in retinal vessel segmentation, we propose a comprehensive micro-vessel extraction mechanism based on an encoder- decoder neural network architecture. This network includes residual, encoder booster, bottleneck enhancement, squeeze, and excitation building blocks. These components synergistically enhance feature extraction and improve the prediction accuracy of the segmentation map. Our solution has been evaluated using the DRIVE, CHASE-DB1, and STARE datasets, yielding competitive results compared to previous studies. The AUC and accuracy on the DRIVE dataset are 0.9884 and 0.9702, respectively. For the CHASE-DB1 dataset, these scores are 0.9903 and 0.9755, respectively, and for the STARE dataset, they are 0.9916 and 0.9750. Given its accurate and robust performance, the proposed approach is a solid candidate for being implemented in real-life diagnostic centers and aiding ophthalmologists.

关键词： Retinal vascular segmentation Microvascular structures encoder booster Bottleneck enhancement encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

A data-driven approach for high accurate spatiotemporal precipitation estimation

引用

NEURAL COMPUTING & APPLICATIONS 2024年第11期36卷 6099-6118页

作者： Pham, Minh Khiem Nguyen, Phi Le Vu, Viet Hung Truong, Thao Nguyen Vo-Van, Hoa Ngo-Duc, Thanh Hanoi Univ Sci & Technol Hanoi Vietnam Natl Inst Adv Ind Sci & Technol Tokyo Japan Northern Delta Reg Hydrometeorol Ctr Hanoi Vietnam Univ Sci & Technol Hanoi USTH Vietnam Acad Sci & Technol VAST Hanoi Vietnam

Precipitation is a fundamental factor affecting many fields, including freshwater reservation, flood warning and prevention, agriculture, and hydropower planning. Observation-based precipitation data usually come from two primary sources, namely gauges and satellite images. The former provides high reliability but sparse coverage, while the latter offers fine-grained data but is still inaccurate. There have been several efforts to estimate precipitation, including mathematical and probabilistic models and machine learning-based techniques. All existing solutions consider the target problem as a satellite data calibration with gauge data as complementary information. However, this approach fails to provide accurate predictive results due to the sparsity of gauges and significant errors in satellite data. This paper presents a novel precipitation estimating method that highlights the importance of gauge data, the most trustworthy data source. To be more precise, we formulate the precipitation estimation issue as a spatial prediction task with the goal of predicting rainfall data for non-monitoring locations using gauge data at the monitored sites. To this end, we propose a data-driven approach that exploits the encoder-decoder architecture, graph neural network, and the multimodal data fusion strategy. Specifically, we design an encoder that leverages a graph neural network for capturing the spatial relationship among the gauges. Meanwhile, the decoder exploits the convolutional networks to learn the temporal correlation within the historical data. Finally, we integrate satellite images and meteorological information using a multimodal data fusion based on a multilayer perceptron to enhance prediction accuracy. The experimental results show that our proposed model increases the estimation accuracy from 24.3 to 65.2% compared to the state-of-the-art.

关键词： Precipitation estimation Spatiotemporal prediction Graph neural networks encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

Learning distribution of video captions using conditional GAN

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2024年第3期83卷 9137-9159页

作者： Babavalian, Mohammad Reza Kiani, Kourosh Semnan Univ Fac Elect & Comp Engn Semnan Iran

Automatic video captioning aims to generate captions with textual descriptions to express video content in natural language by a machine. This is a difficult task as the videos contain dynamic challenges. Most of the available approaches for video captioning are often focused on providing a single descriptive sentence. encoder-decoder is the most popular architecture developed for video captioning. The proposed method in this research aims to learn the distribution of captions to generate more relevant and diverse captions and increase generalizability. A novel architecture was developed based on conditional SeqGAN to learn the distribution for video captioning and increase the generalizability. This architecture consists of two modules: encoding and caption generation. The goal of encoding is to obtain encoded rich spatial-temporal features. The encoding vector is fed as the input of conditional SeqGAN to generate captions. The main novelty of this paper lies in the use of an adversarial approach to learn the distribution of captions and generate diverse captions that fit the characteristics of the video. Experimental results from two popular datasets, MSVD and MSRVTT, showed that the proposed approach achieved more relevant video captions than other state-of-the-art methods.

关键词： Video captioning Captions distribution Conditional GAN encoder-decoder Captions generation Adversarial Video Captioning

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：