In order to improve the efficiency and visual performance of film production, this article analyzes virtual production technology based on deep learning, especially in areas such as scene reconstruction, dynamic captu...
详细信息
ISBN:
(纸本)9798400707032
In order to improve the efficiency and visual performance of film production, this article analyzes virtual production technology based on deep learning, especially in areas such as scene reconstruction, dynamic capture, and special effects generation. The results indicate that convolutional neural networks and generative adversarial networks have significant advantages in imageprocessing and virtual content generation, while long short-term memory networks perform well in dynamic scene generation, demonstrating the widespread potential of deep learning technology in virtual film production.
The security of image information is becoming a growing concern, image encryption technology is becoming more and more comprehensive. The algorithm encrypts the image chunks, decoding and operation mode of each image ...
详细信息
ISBN:
(纸本)9798400707032
The security of image information is becoming a growing concern, image encryption technology is becoming more and more comprehensive. The algorithm encrypts the image chunks, decoding and operation mode of each image block. It cannot resist clipping attacks, this paper increases the number of chaotic systems in the algorithm and optimizes the encryption process. The results show that the key capacity of the improved encryption algorithm is increased to 10(127), which can resist the exhaustive key attack. After encryption, the correlation of adjacent elements of the image is reduced to 10(3) orders of magnitude, and The encrypted image is crop-resistant.
The accuracy of the output data of inertial navigation system is closely related to temperature, so how to ensure the accuracy, stability and reliability of temperature acquisition is very important for inertial navig...
详细信息
ISBN:
(纸本)9798400707032
The accuracy of the output data of inertial navigation system is closely related to temperature, so how to ensure the accuracy, stability and reliability of temperature acquisition is very important for inertial navigation system. Based on engineering practice and analysis and comparison of common temperature processing software methods, this paper proposes a smoothing algorithm for establishing temperature reference, which can eliminate possible extreme values and ensure that the temperature change rate is within the expected range, further reduce temperature fluctuations, and improve the accuracy of navigation operations and the stability of subsequent navigation operations.
In recent years, with the rapid development of artificial intelligence, multi-modal knowledge graph completion (MMKGC) has become increasingly important. Many scholars have conducted in-depth research on multi-modal k...
详细信息
ISBN:
(纸本)9798400707032
In recent years, with the rapid development of artificial intelligence, multi-modal knowledge graph completion (MMKGC) has become increasingly important. Many scholars have conducted in-depth research on multi-modal knowledge graphs (MMKGs), leading to the proposal of numerous MMKGC models. Summarizing the current state of research is crucial for guiding future studies. This survey aims to review the current advanced techniques for MMKGC. By analyzing and elaborating on the value and categories of MMKGs in detail, we summarize the challenges faced by existing MMKGC methods. Our work provides valuable insights and explorations for the research and application of completing MMKGs.
A method for detecting defects in forming mesh based on the YOLOV8N model is proposed. Firstly, an illumination system is designed, and then image data with defects is collected to construct an image dataset. The cons...
详细信息
ISBN:
(纸本)9798400707032
A method for detecting defects in forming mesh based on the YOLOV8N model is proposed. Firstly, an illumination system is designed, and then image data with defects is collected to construct an image dataset. The constructed image dataset is annotated and augmented to build a corresponding image sample set. The YOLOV8N model is trained using the sample set to obtain a defect recognition and localization model. The original image to be recognized is input into the defect recognition and localization model, which then outputs the corresponding defect recognition and localization results. Experimental results show that this method provides accurate localization and fast speed.
In this study, a deep learning-based galaxy classification method is discussed. The classification of galaxies is important for understanding the formation and evolution of the universe, and traditional classification...
详细信息
ISBN:
(纸本)9798400707032
In this study, a deep learning-based galaxy classification method is discussed. The classification of galaxies is important for understanding the formation and evolution of the universe, and traditional classification methods rely on artificial visual analysis, but in the face of large amounts of data, this method is time-consuming and prone to error. In recent years, automated classification methods, especially using deep learning techniques, have gradually come into focus. Deep learning models, particularly convolutional neural networks (CNNS), are capable of automatically extracting and learning complex features in galactic images, enabling efficient and accurate classification. The research plan is to integrate multimodal data, train models with large-scale datasets, and introduce interpretative analysis into the classification process to improve model transparency. Ultimately, the goal is to develop an efficient galactic classification system to support data processing and analysis in the field of astronomy.
To address the issues of limited samples, time-consuming feature design, and low accuracy in detection and classification of breast cancer pathological images, a breast cancer image classification model algorithm comb...
详细信息
ISBN:
(纸本)9798400707032
To address the issues of limited samples, time-consuming feature design, and low accuracy in detection and classification of breast cancer pathological images, a breast cancer image classification model algorithm combining deep learning and transfer learning is proposed. This algorithm is based on the Densely Connected Convolutional Networks (DenseNet) structure of deep neural networks, and constructs a network model by introducing attention mechanisms, and trains the enhanced dataset using multi-level transfer learning. Experimental results demonstrate that the algorithm achieves an efficiency of over 84.0% in the test set, with a significantly improved classification accuracy compared to previous models, making it applicable to medical breast cancer detection tasks.
Participants in the supply chain may have different information, leading to incomplete or inaccurate information when making decisions. To this end, a process and machinelearning based collaborative scheduling algori...
详细信息
ISBN:
(纸本)9798400707032
Participants in the supply chain may have different information, leading to incomplete or inaccurate information when making decisions. To this end, a process and machinelearning based collaborative scheduling algorithm for all materials is proposed. Design a health monitoring process for material supply chain based on R-tree dynamic indexing algorithm. Based on this, artificial neural networks in machinelearning are applied to mine the data of the entire material supply chain. Through data mining, various data in the supply chain can be integrated and analyzed to improve information transparency and accuracy, and reduce information asymmetry. Adopting a dual layer scheduling model to achieve dual layer collaborative scheduling of materials. The experimental results show that the research method effectively improves the accuracy of data mining in the entire material supply chain, and the utilization rate of materials under this method is always higher than 95%.
A statistical correlation model for image retrieval is proposed. This model captures the semantic relationships among images in a database from simple statistics of user-provided relevance feedback information. It is ...
详细信息
A statistical correlation model for image retrieval is proposed. This model captures the semantic relationships among images in a database from simple statistics of user-provided relevance feedback information. It is applied in the post-processing of image retrieval results such that more semantically related images are returned to the user. The algorithm is easy to implement and can be efficiently integrated into an image retrieval system to help improve the retrieval performance. Preliminary experimental results on a database of 100,000 images show that the proposed model could improve image retrieval performance for both content- and text-based queries. (C) 2002 patternrecognition Society. Published by Elsevier Science Ltd. All rights reserved.
In the realm of contemporary deep learning, the pivotal challenges confronting few-shot and unsupervised learning reside in harnessing scarce labeled samples alongside abundant unlabeled ones, particularly in the cont...
详细信息
ISBN:
(纸本)9798400707032
In the realm of contemporary deep learning, the pivotal challenges confronting few-shot and unsupervised learning reside in harnessing scarce labeled samples alongside abundant unlabeled ones, particularly in the context of medical image classification. This paper introduces an innovative approach that seamlessly integrates dynamic clustering with weights augmentation, aimed at bolstering the performance of few-shot medical image classification. Dubbed Weight-Enhanced Contrastive learning (WECL), our method ingeniously fuses contrastive representation learning with a dynamic memory module during unsupervised pre-training. This fusion facilitates efficient clustering and classification of diverse augmented renditions of the same image. Additionally, the weights augmentation tactic meticulously tunes the weights of both ResNet and teacher-student model branches, thereby mitigating sample bias and enhancing the pre-trained model's proficiency. Extensive experiments across multiple few-shot medical image classification datasets underscore the superiority of our WECL approach, outperforming current state-of-the-art baselines, and effectively addressing issues pertaining to data distribution disparities and sample scarcity.
暂无评论