检索结果-内蒙古大学图书馆

Symmetrical filters in convolutional neural networks

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 2021年第7期12卷 2027-2039页

作者： Dzhezyan, Gregory Cecotti, Hubert Fresno State Dept Comp Sci Coll Sci & Math Fresno CA 93740 USA

Symmetry is present in nature and science. In image processing, kernels for spatial filtering possess some symmetry (e.g. Sobel operators, Gaussian, Laplacian). Convolutional layers in artificial feed-forward neural networks have typically considered the kernel weights without any constraint. We propose to investigate the impact of a symmetry constraint in convolutional layers for image classification tasks, taking our inspiration from the processes involved in the primary visual cortex and common image processing techniques. The goal is to determine if it is necessary to learn each weight of the filters independently, and the extent to which it is possible to enforce symmetrical constraints on the filters throughout the training process of a convolutional neural network by modifying the weight update preformed during the backpropagation algorithm and to evaluate the change in performance. The symmetrical constraint reduces the number of free parameters in the network, and it is able to achieve near identical performance. We address the following cases: x/y-axis symmetry, point reflection, and anti-point reflection. The performance is evaluated on four databases of images representing handwritten digits. The results support the conclusion that while random weights offer more freedom to the model, the symmetry constraint provides a similar level of performance while decreasing substantially the number of free parameters in the model. Such an approach can be valuable in phase-sensitive applications that require a linear phase property throughout the feature extraction process.

关键词： Convolutional neural network Symmetry Deep learning image processing

来源：评论

学校读者我要写书评

暂无评论

Realistic and Visually-Pleasing 3D Generation of Indoor Scenes from a Single image 7th

Realistic and Visually-Pleasing 3D Generation of Indoor Scen...

引用

7th Chinese Conference on Pattern Recognition and Computer Vision

作者： Li, Jie Wang, Lei Chen, Gongbin Li, Ang Qiu, Yuhao Wu, Jiaji Cheng, Jun Chinese Acad Sci Shenzhen Inst Adv Technol Shenzhen 518055 Peoples R China Univ Chinese Acad Sci CAS Beijing Peoples R China Shenzhen MSU BIT Univ Shenzhen Peoples R China Xidian Univ Sch Elect Engn Xian 710071 Peoples R China

ISBN: (纸本)9789819785070;9789819785087

artificial Intelligence Generated Content (AIGC) has experienced significant advancements, particularly in the areas of natural language processing and 2D image generation. However, the generation of three-dimensional (3D) content from a single image still poses challenges, particularly when the input image contains complex backgrounds. This limitation hinders the potential applications of AIGC in areas such as human-machine interaction, virtual reality (VR), and architectural design. Despite the progress made so far, existing methods face difficulties when dealing with single images that have intricate backgrounds. Their reconstructed 3D shapes tend to be incomplete, noisy, or lack of partial geometric structures. In this paper, we introduce a 3D generation framework for indoor scenes from a single image to generate realistic and visually-pleasing 3D geometry shapes, without the requirement of point clouds, multi-view images, depth or masks as input. The main idea of our method is clustering-based 3D shape learning and prediction, followed by a shape deformation. Since more than one objects tend to be existing in indoor scenes, our framework will simultaneously generate multi-objects and predict the layout with a camera pose, as well as 3D object bounding boxes for holistic 3D scene understanding. We have evaluated the proposed framework on benchmark datasets including ShapeNet, SUN RGB-D and Pix3D, and state-of-the-art performance has been achieved. We have also given examples to illustrate immediate applications in virtual reality.

关键词： 3D mesh Reconstruction Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

New Optimization: Design of a neural Network Cat and Dog Recognition Prediction Model 5

New Optimization: Design of a Neural Network Cat and Dog Rec...

引用

2023 5th International Conference on artificial Intelligence and Computer applications, ICAICA 2023

作者： Zhang, Xinwei Zhang, Longqing Song, Zi'ang Guangdong University of Science and Technology Dongguan China Hebei University of Environmental Engineering Qinhuangdao China

ISBN: (纸本)9798350323313

With the advent of the big data era, there has been a surge in research focused on the application of convolutional neural networks (CNNs) and image processing. Similar to how humans effortlessly identify cats and dogs in our everyday lives, modern machines are also capable of performing this task. One effective approach to achieving this is through leveraging Kaggle's automatic cat and dog identification technology. This study aims to tackle the challenge of recognizing cats and dogs in images with complex appearances by developing a Sequential object-based CNN model. To accomplish this, the cutting-edge deep learning framework PyTorch is utilized, along with high-performance GPUs for computational power. The network is trained and tested separately on images of dogs and cats. Experimental results demonstrate that the CNN model achieves remarkably high recognition accuracy and exhibits exceptional performance in distinguishing between different breeds of dogs and cats. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Anchor-free object detection in surveillance applications

Anchor-free object detection in surveillance applications

引用

作者： Magnusson, Peter Malmo University

学位级别：硕士

Computer vision object detection is the task of detecting and identifying objects present in an image or a video sequence. Models based on artificial convolutional neural networks are commonly used as detector models. Object detection precision and inference efficiency are crucial for surveillance-based applications. A decrease in the detector model complexity as well as in the complexity of the post-processing computations promotes increased inference efficiency. Modern object detectors for surveillance applications usually make use of a regression algorithm and bounding box priors referred to as anchor boxes to compute bounding box proposals, and the proposal selection algorithm contributes to the computational cost at inference. In this study, an anchor-free and low complexity deep learning detector model was implemented within a surveillance applications setting, and was evaluated and compared to a reference baseline state-of-the-art anchor-based object detector. A key-point-based detector model (CenterNet), predicting Gaussian distribution based object centers, was selected for the evaluation against the baseline. The surveillance applications adapted anchor-free detector exhibited a factor 2.4 lower complexity than the baseline detector. Further, a significant redistribution to shorter post-processing times was demonstrated at inference for the anchor-free surveillance adapted CenterNet detector, giving a modal values factor 0.6 of the baseline detector post-processing time. Furthermore, the surveillance adapted CenterNet model was shown to outperform the baseline in terms of detection precision for several surveillance applications relevant classes and for objects of smaller spatial scale.

关键词：

来源：评论

学校读者我要写书评

暂无评论

On the Improvement of Generalization and Stability of Forward-Only Learning via neural Polarization 27

On the Improvement of Generalization and Stability of Forwar...

引用

27th European Conference on artificial Intelligence, ECAI 2024

作者： Terres-Escudero, Erik B. Del Ser, Javier Garcia-Bringas, Pablo University of Deusto Bilbao48007 Spain Derio48160 Spain Bilbao48013 Spain

ISBN: (纸本)9781643685489

Forward-only learning algorithms have recently gained attention as alternatives to gradient backpropagation, replacing the backward step of this latter solver with an additional contrastive forward pass. Among these approaches, the so-called Forward-Forward Algorithm (FFA) has been shown to achieve competitive levels of performance in terms of generalization and complexity. networks trained using FFA learn to contrastively maximize a layer-wise defined goodness score when presented with real data (denoted as positive samples) and to minimize it when processing synthetic data (corr. negative samples). However, this algorithm still faces weaknesses that negatively affect the model accuracy and training stability, primarily due to a gradient imbalance between positive and negative samples. To overcome this issue, in this work we propose a novel implementation of the FFA algorithm, denoted as Polar-FFA, which extends the original formulation by introducing a neural division (polarization) between positive and negative instances. Neurons in each of these groups aim to maximize their goodness when presented with their respective data type, thereby creating a symmetric gradient behavior. To empirically gauge the improved learning capabilities of our proposed Polar-FFA, we perform several systematic experiments using different activation and goodness functions over image classification datasets. Our results demonstrate that Polar-FFA outperforms FFA in terms of accuracy and convergence speed. Furthermore, its lower reliance on hyperparameters reduces the need for hyperparameter tuning to guarantee optimal generalization capabilities, thereby allowing for a broader range of neural network configurations. © 2024 The Authors.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Video anomaly detection with autoregressive modeling of covariance features

引用

SIGNAL image AND VIDEO processing 2022年第4期16卷 1027-1034页

作者： Bilecen, Ali Enver Ozalp, Alp Yavuz, M. Sami Ozkan, Huseyin Sabanci Univ Fac Engn & Nat Sci Istanbul Turkey

In this paper, we propose three different methods for anomaly detection in surveillance videos based on modeling of observation likelihoods. By means of the methods we propose, normal (typical) events in a scene are learned in a probabilistic framework by estimating the features of consecutive frames taken from the surveillance camera. The proposed methods are based on long short-term memory (LSTM) and linear regression. To decide whether an observation sequence (i.e., a small video patch) contains an anomaly or not, its likelihood under the modeled typical observation distribution is thresholded. An anomaly is decided to be present if the threshold is exceeded. Due to its effectiveness in object detection and action recognition applications, covariance features are used in this study to compactly reduce the dimensionality of the shape and motion cues of spatiotemporal patches obtained from the video segments. The two most successful methods are based on the final state vector of LSTM and support vector regression applied to mean covariance features and achieve an average performance of up to 0.95 area under curve on benchmark datasets.

关键词： Anomaly detection Covariance features LSTM Linear regression artificial neural networks

来源：评论

学校读者我要写书评

暂无评论

Assessment algorithm of power supply reliability for customers in rural healthcare centres

引用

Journal of Commercial Biotechnology 2022年第5期27卷 239-252页

作者： Wei, Zhifeng Gao, Bingqiang Liu, Haiyang An, Lili Wang, Pengfei Xia, Liu Fei, Changshun Beijing State Grid ICT Accenture Information Technology Co. Beijing100032 China

Power supply interruptions in low-voltage customers, caused by blackouts and other factors, can significantly impact the functioning of rural healthcare centres. To address this issue, the development of a predictive system that ensures a reliable power supply becomes essential. artificial neural networks (ANNs) emerge as promising predictive systems due to their efficacy in various artificial intelligence and computer science problems, such as pattern recognition and diagnosis. In this study, we explore the application of three different ANN models: Nonlinear Multilayer Perceptron (MLP), Radial Basis Function Network (RBF Net), and Hopfield neural Network (HNN), to enhance the prediction of power supply reliability for low-voltage customers. In addition to ANNs, an optimization algorithm based on Density-Based Spatial Clustering of applications with Noise (DBSCAN) is proposed to identify obstacle hazards. The traditional DBSCAN algorithm's limitations in adapting to unmanned obstacle data processing are addressed through improvements. By adjusting the fixed cluster neighbourhood radius to an adaptive clustering parameter that varies with the target distance, the clustering effect of the optimized DBSCAN algorithm is significantly enhanced. Through experimental verification and analysis, we demonstrate the effectiveness of the ANN models and the improved DBSCAN algorithm in enhancing power supply reliability assessment and obstacle hazard identification. The combination of advanced predictive systems and optimized clustering algorithms promise to ensure an uninterrupted and reliable power supply for rural healthcare centers, improving the quality of healthcare services in underserved areas. © 2022 ThinkBiotech LLC. All rights reserved.

关键词： image processing

来源：评论

学校读者我要写书评

暂无评论

Application Analysis of Digital Special Effects Technology in Film and Television Post-production Based on neural Network Algorithm 4th

Application Analysis of Digital Special Effects Technology i...

引用

4th International Conference on Machine Learning, image processing, Network Security and Data Sciences (MIND)

作者： Qian, Hongxing Changchun Humanities & Sci Coll Sch Media & Commun Changchun Jilin Peoples R China

ISBN: (纸本)9783031243660;9783031243677

The leap in film and television special effects technology has updated a series of film production methods. Post-production using the most advanced computer graphics technology stimulates the creativity of the producers, simplifies the post-production process, and improves the quality of the entire film. The purpose of this paper is to study the application of digital special effects technology in film and television post-production based on neural network algorithm. First, the digital technology used in the widely used film and television post-production is introduced, and then some applications and problems of artificial neural networks are introduced. Then the PointNet network structure is introduced, which is a deep learning network in 3D point cloud. Framework, which eliminates the ambiguity caused by the disorder and rotation of point clouds by introducing T-net and utilizing max-pooling, and finally we introduce an encoder-decoder network for 3D human reconstruction, which encodes The network extracts the features, and uses the decoding network to learn the transformation between the template and the input point cloud, so as to complete the deformation fitting between the template and the point cloud.

关键词： neural networks Film and television Post production Digital special effects technology

来源：评论

学校读者我要写书评

暂无评论

Optics, Photonics and Digital Technologies for Imaging applications vii

Optics, Photonics and Digital Technologies for Imaging Appli...

引用

Optics, Photonics and Digital Technologies for Imaging applications vii 2022

ISBN: (纸本)9781510651524

The proceedings contain 35 papers. The topics discussed include: noise robust focal distance detection in laser material processing using CNNs and Gaussian processes;machine learning-based high-precision and real-time focus detection for laser material processing systems;sargassum detection and path estimation using neural networks;neuron segmentation in epifluorescence microscopy imaging with deep learning;multimodal super-resolution reconstruction based on encoder-decoder network;synthetic apertures for array ptychography imaging via deep learning;infrared image super-resolution pseudo-color reconstruction based on dual-path propagation;effective laser pest control with modulated UV-A light trapping for mushroom fungus gnats;and integration of augmented reality and image processing in plasma dynamic analysis: digital concepts and structural system design.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lightweight Modules for Efficient Deep Learning Based image Restoration

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2021年第4期31卷 1395-1410页

作者： Lahiri, Avisek Bairagya, Sourav Bera, Sutanu Haldar, Siddhant Biswas, Prabir Kumar Indian Inst Technol Kharagpur Kharagpur 721302 W Bengal India Mathworks New Delhi 110001 India

Low level image restoration is an integral component of modern artificial intelligence (AI) driven camera pipelines. Most of these frameworks are based on deep neural networks which present a massive computational overhead on resource constrained platform like a mobile phone. In this paper, we propose several lightweight low-level modules which can be used to create a computationally low cost variant of a given baseline model. Recent works for efficient neural networks design have mainly focused on classification. However, low-level image processing falls under the 'image-to-image' translation genre which requires some additional computational modules not present in classification. This paper seeks to bridge this gap by designing generic efficient modules which can replace essential components used in contemporary deep learning based image restoration networks. We also present and analyse our results highlighting the drawbacks of applying depthwise separable convolutional kernel (a popular method for efficient classification network) for sub-pixel convolution based upsampling (a popular upsampling strategy for low-level vision applications). This shows that concepts from domain of classification cannot always be seamlessly integrated into 'image-to-image' translation tasks. We extensively validate our findings on three popular tasks of image inpainting, denoising and super-resolution. Our results show that proposed networks consistently output visually similar reconstructions compared to full capacity baselines with significant reduction of parameters, memory footprint and execution speeds on contemporary mobile devices.

关键词： Convolution image restoration Task analysis neural networks Kernel Computational modeling image denoising image inpainting image super-resolution CNN generative adversarial network (GAN) adversarial learning efficient neural networks

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：