Accurately predicting aerodrome traffic pattern trajectories is a crucial and challenging task in aviation operation optimization. To address the problem that the currently commonly used flight trajectory prediction m...
详细信息
Stereo matching is an essential basis for various applications in computervision, but currently, most stereo matching methods have poor generalization performance and require a fixed disparity search range. In this w...
详细信息
ISBN:
(数字)9781665490627
ISBN:
(纸本)9781665490627
Stereo matching is an essential basis for various applications in computervision, but currently, most stereo matching methods have poor generalization performance and require a fixed disparity search range. In this work, we propose to adopt the operation "sub-pixel three-around-maximum", which supports an unfixed disparity search range, to replace the currently popular operation soft argmax. We also propose to directly supervise the feature extractor by three loss functions. Just depending on the feature extractor, we can obtain an accurate semi-dense disparity map before cost aggregation, which can help to remove adaptively the redundancy part of the predefined disparity search range. An adaptive disparity search range for each stereo pair can save much time and memory. the proposed architecture achieves state-of-the-art cross-domain generalization performance in the datasets KITTI2012&2015, and experimental results demonstrate that our method supports unfixed and adaptive disparity search range.
the proceedings contain 38 papers. the special focus in this conference is on Wireless Communications, Networking and Applications. the topics include: EffiNet: An Efficient, Low-Complexity Neural Network for Mon...
ISBN:
(纸本)9789819624089
the proceedings contain 38 papers. the special focus in this conference is on Wireless Communications, Networking and Applications. the topics include: EffiNet: An Efficient, Low-Complexity Neural Network for Monocular Depth Estimation on Embedded Devices;adaptive Communication Spectrum Sensing Algorithm Based on Energy Detection;a Cache Scheduling Method Based on Adaptive Expiration for Data Process System;Research on Airport Communication and Navigation Equipment Inspection Assistance System Based on AR Technology;application of Quality Function Deployment in Assessing Unmanned Assault Capabilities for Mountain Counter-Terrorism Operations;research on the Application of Internet of things Technology in Power Grid Limited Space Security Monitoring;research on the Application of computervision Based on Deep Learning in Autonomous Driving Technology;research on Cybersecurity System of Scientific Research Institutions Based on National Cybersecurity Classified Protection 2.0: Take the chinese Academy of Agricultural Sciences as an Example;Research on DOA Estimation Method of Underwater Acoustic Signal Based on Machine Learning;Design and Implementation of Smart Home Control System Based on STM32;communication Anti-jamming System Based on Deep Reinforcement Learning;fault Diagnosis of Infrared Sensor Based on Convolutional Neural Network;graph Neural Network Knowledge Graph Recommendation Model Integrating Deep Domain Information and Important Domain Information;an Approach to Microenvironment-Based Particle Swarm Optimization Algorithm;a Verifiable Ciphertext Retrieval Scheme for Smart Grids;panoptic Semantic Mapping Method for Tomato Growing Environment Based on K-Net and OctoMap;An Adversarial Attack Method for Multivariate Time Series Classification Based on AdvGAN;recognition and Calculation of Fish Rafts in Mariculture on the Basis of Artificial Intelligence;digital Twin for Power Load Forecasting.
Text-line segmentation is still considered challenging for complex background scene images. the success of text detection and recognition depends on the success of the text segmentation. this study presents a new meth...
详细信息
In this paper, we focus on a critical task of retrieving common style in chinese scene text: given an image of style text, the system returns all the images matching the queried text image. To that, a novel twin Trans...
详细信息
ISBN:
(纸本)9783031189128;9783031189135
In this paper, we focus on a critical task of retrieving common style in chinese scene text: given an image of style text, the system returns all the images matching the queried text image. To that, a novel twin Transformer based matching network is proposed, which is featured by the integration of anchor-free detection, text recognition, and similarity matching networks. On the fly, our model retrieves the similarity of text features in the text area and evaluates it through recognition. Our experiments demonstrate that the proposed model outperforms the state-of-the-art in terms of both processing speed and accuracy. Additional experiments show that our model generalizes well on various benchmarks, including a self-constructed chinese query data set with complex chinese scenes in the real world.
Retrieving tracked-vehicles by natural language descriptions plays a critical role in smart city construction. It aims to find the best match for the given texts from a set of tracked vehicles in surveillance videos. ...
详细信息
ISBN:
(数字)9781665487399
ISBN:
(纸本)9781665487399
Retrieving tracked-vehicles by natural language descriptions plays a critical role in smart city construction. It aims to find the best match for the given texts from a set of tracked vehicles in surveillance videos. Existing works generally solve it by a dual-stream framework, which consists of a text encoder, a visual encoder and a cross-modal loss function. Although some progress has been made, they failed to fully exploit the information at various levels of granularity. To tackle this issue, we propose a novel framework for the natural language-based vehicle retrieval task, OMG, which Observes Multiple Granularities with respect to visual representation, textual representation and objective functions. For the visual representation, target features, context features and motion features are encoded separately. For the textual representation, one global embedding, three local embeddings and a color-type prompt embedding are extracted to represent various granularities of semantic features. Finally, the overall framework is optimized by a cross-modal multi-granularity contrastive loss function. Experiments demonstrate the effectiveness of our method. Our OMG significantly outperforms all previous methods and ranks the 9th on the 6th AI City Challenge Track2.
Specular highlight detection and removal are fundamental challenges in computervision and image processing, withthe detection results serving as a precursor to guide the model in achieving better removal of specular...
详细信息
ISBN:
(纸本)9783031705328;9783031705335
Specular highlight detection and removal are fundamental challenges in computervision and image processing, withthe detection results serving as a precursor to guide the model in achieving better removal of specular highlights. this paper introduces a novel highlight removal model, which presents an efficient end-to-end deep learning framework designed to automatically remove specular highlights from a single image. Our architecture comprises three key modules: the Coarse Predictor (CP), Refinement Predictor (RP), and Global Discriminator (GD). the CP utilizes a novel Transformer-based Unet architecture to recover the primary content, while the GD incorporates a discriminator to ensure the coarse result is more feasible in a global context. Lastly, the RP is based on conditional Denoising Diffusion Probabilistic Models (DDPM) and is responsible for predicting the residual information between the ground-truth and the CP-predicted image. Experimental results on four public benchmark images demonstrate that our method surpasses state-of-the-art methods in the task of highlight removal.
3D human pose estimation is pivotal in the domains of computervision and kinematic analysis. However, the majority of existing 3D datasets focus on indoor environments, resulting in a significant gap in extensive res...
详细信息
the limited wind power output caused by cold wave weather poses challenges to the safe and stable operation of the power system. In order to improve the accuracy of wind power output recognition under cold weather con...
详细信息
Textile patternrecognition was studied based on artificial intelligence (AI) for the automation of patternrecognition on textiles. computervision and deep learning techniques were used for this study. Textile image...
Textile patternrecognition was studied based on artificial intelligence (AI) for the automation of patternrecognition on textiles. computervision and deep learning techniques were used for this study. Textile images of different textiles were collected as data. the data was preprocessed to extract features to construct a dataset for patternrecognition. Subsequently, traditional machine learning algorithms and deep learning methods were used to establish a classification model for textile patterns. the performance of different models was evaluated by comparison. In the experiments, the robustness of the model was tested for diversity and variability. the model was trained in the actual textile production line to test for the real-time recognition of textile patterns. the model can be used for automated production in the textile industry with a wide range of application prospects.
暂无评论