The bone symmetry and motion locality are two important priors for 3D human pose estimation. For bone symmetry, symmetric bones typically have equal length and higher motion correlation. In terms of motion locality, t...
详细信息
ISBN:
(数字)9798350369151
ISBN:
(纸本)9798350369168
The bone symmetry and motion locality are two important priors for 3D human pose estimation. For bone symmetry, symmetric bones typically have equal length and higher motion correlation. In terms of motion locality, the motion of the human body conforms to its structure and is temporally continuous. These two priors align with human intuition, but are challenging for Transformer to directly abstract them from data. We propose novel modules, Bone Symmetry Temporal Transformer (BSTT) and Spatial-Temporal Locality Fusion (STLF), to utilize these two priors, and design a new Transformer-based BSMLFormer model. Specifically, BSTT considers the temporal sequences of two symmetric bones as a single new sequence and calculates its global relationships by self-attention mechanism. STLF performs local computations by GCN and convolutions to capture the local dependencies in adjacent joints and adjacent parts. The local-global alternating structure of BSMLFormer allows for consideration of the local information during global calculations. More remarkably, BSMLFormer achieves the state-of-the-art performance with 39.0mm MPJPE on Human3.6M dataset.
Computer Vision has given a way for computer systems to see by way of deciphering the surrounding objects, which has been considered a crucial problem over the years. There are quite a number of techniques devised for...
详细信息
Computer Vision has given a way for computer systems to see by way of deciphering the surrounding objects, which has been considered a crucial problem over the years. There are quite a number of techniques devised for object detection and deep learning, but most of the research has been focused on deep learning in recent years. Visual object detection covers numerous recognitionpattern tasks like image classification. This article aims to review a visual analytics approach for better understanding, identifying, and cleansing object detection frameworks and the effects of some factors such as sampling strategies, feature learning, detector architectures, proposal generation, etc., on visual object detection with special reference to detection components, learning strategies, and their applications along with benchmarks.
Cobb angle on X-ray image is a gold standard for the evaluation of scoliosis. When measuring the Cobb angle, there are some artificial measurement errors, which lead to the unreliable results. A spine curve extraction...
详细信息
ISBN:
(纸本)9781450390057
Cobb angle on X-ray image is a gold standard for the evaluation of scoliosis. When measuring the Cobb angle, there are some artificial measurement errors, which lead to the unreliable results. A spine curve extraction method based on mask segmentation is proposed in this paper. The central angle of the arc corresponding to the main C bending of the spine is calculated to evaluate the degree of scoliosis. Select 173 X-ray digital images of the patients with scoliosis randomly. Firstly, the two spine curves are extracted by using the new method proposed in this paper and the traditional cross method separately, and the central angles corresponding to the main C bending of the two different spine curves mentioned above are measured by three experienced surveyors respectively. After a week, the sets of data were measured repeatedly. SPSS 22.0 statistical analysis software is used to analyze the results statistically. Compared with the spine curve extracted by the cross method, the new method proposed in this paper has better reliability and stability in evaluating the degree of scoliosis.
Transformer and Convolutional Neural Network (CNN) are currently two important models in the field of deep learning. Among them, Transformer has strong global perception ability but weak local perception ability, and ...
详细信息
ISBN:
(数字)9798350365443
ISBN:
(纸本)9798350365450
Transformer and Convolutional Neural Network (CNN) are currently two important models in the field of deep learning. Among them, Transformer has strong global perception ability but weak local perception ability, and CNN precisely compensates for this. CNN has strong local perception ability, so in order to combine the advantages of both, we propose a 3D classification model TCPNet based on Transformer and CNN. We first use Transformer to extract the global features of the point cloud, and then concatenate a CNN network to enhance the detail perception ability, so that the network can better understand the 3D point cloud. The experimental results on the ModelNet40 and ScanObjectNN datasets show that our method can effectively achieve 3D classification tasks. On ModelNet40, an accuracy of 93.6% can be achieved. On the most difficult variant PB-T50-RS of ScanObjectNN, an accuracy of 83.93% can be achieved.
This Study would identify regional fabric patterns throughout Indonesia. Traditional clothes patterns come from the handicrafts of remote communities in the regions of Indonesia. A Convolutional neural network is used...
详细信息
The modern world healthy body depends on the number of calories consumed, hence monitoring calorie intake is necessary to maintain good health. At the point when your Body Mass Index is somewhere in between from 25 to...
详细信息
ISBN:
(纸本)9781665428644
The modern world healthy body depends on the number of calories consumed, hence monitoring calorie intake is necessary to maintain good health. At the point when your Body Mass Index is somewhere in between from 25 to 29. It implies that you are conveying overabundance weight. Assuming your BMI is more than 30, it implies you have obesity. To get in shape or keep up the solid weight individuals needs to monitor the calorie they take. The existing system calorie estimation is to be happened manually. The proposed model is to provide unique solution for measuring calorie by using deep learning algorithm. The food calorie calculation is very important in medical field. Because this food calorie is provide good health condition. This measurement is taken from food image in different objects that is fruits and vegetables. This measurement is taken with the help of neural network. The tensor flow is one of the best methods to classify the machine learning method. This method is implementing to calculate the food calorie with the help of Convolutional Neural Network. The input of this calculated model is taken an image of food. The food calorie value is calculated the proposed CNN model with the help of food object detection. The primary parameter of the result is taken by volume error estimation and secondary parameter is calorie error estimation. The volume error estimation is gradually reduced by 20%. That indicates the proposed CNN model is providing higher accuracy level compare to existing model.
Optical Character recognition (OCR) is the most successful application for automatic patternrecognition. OCR has been a widely studied field of research and development for more than last 50 years. It is one of the m...
详细信息
Optical Character recognition (OCR) is the most successful application for automatic patternrecognition. OCR has been a widely studied field of research and development for more than last 50 years. It is one of the many important gifts given by computer science to the mankind. OCR means "a technique of recognition of machine printed or hand written text by computer and its conversion to an editable form as per the requirement". During the recent decades, handwritten character recognition has been a subject of study because of both its theoretical importance in recognition system with numerous possible applications. This paper presents an extensive study of numerous publications by a number of researchers working for years in the field of Odia character recognition (one of the many regional languages used in India and spoken by more than 30 million people worldwide). The main focus has been on feature extraction, data set size, classification techniques and recognition accuracy. A number of selective articles published in journals and conference proceedings on character recognition have been included in this work (2005 to 2020). The review process has been divided into two parts. In the first part a study of various works by researchers is presented with detailed description. In the second part, various feature extraction and classification techniques have been presented and summarized. The findings are presented in the analysis section. (C) 2019 Elsevier Ltd. All rights reserved.
The proceedings contain 13 papers. The special focus in this conference is on Benchmarking, Measuring, and Optimizing. The topics include: Artemis: An Automatic Test Suite Generator for Large Scale OLAP Database;OStor...
ISBN:
(纸本)9783030710576
The proceedings contain 13 papers. The special focus in this conference is on Benchmarking, Measuring, and Optimizing. The topics include: Artemis: An Automatic Test Suite Generator for Large Scale OLAP Database;OStoreBench: Benchmarking Distributed Object Storage Systems Using Real-World Application Scenarios;ConfAdvisor: An Automatic Configuration Tuning Framework for NoSQL Database Benchmarking with a Black-box Approach;Optimization of the Himeno Benchmark for SX-Aurora TSUBASA;Parallel Sorted Sparse Approximate Inverse Preconditioning Algorithm on GPU;preface;characterizing the Sharing Behavior of Applications Using Software Transactional Memory;ComScribe: Identifying Intra-node GPU Communication;a Benchmark of Ocular Disease Intelligent recognition: One Shot for Multi-disease Detection;MAS3K: An Open Dataset for Marine Animal Segmentation;benchmarking Blockchain Interactions in Mobile Edge Cloud Software Systems.
Industrial processes are becoming increasingly complex and larger in scale, which has resulted in a constantly growing demand for robust, scalable monitoring systems. Conventional monitoring strategies entail a huge q...
详细信息
ISBN:
(数字)9798331519582
ISBN:
(纸本)9798331519599
Industrial processes are becoming increasingly complex and larger in scale, which has resulted in a constantly growing demand for robust, scalable monitoring systems. Conventional monitoring strategies entail a huge quantity of sensors and steep infrastructure charges, main to excessive electricity intake and operational costs. Burkhard Theen, vice president at weather analytics company Iteris describes a smart monitoring device using IoT sensors and predictive algorithms for accurately tracking industrial processes on this track. The system is made using a low-voltage IoT sensor community at essential points within the process, which constantly collects and transmits statistics in real time. This data is then fed into the predictive algorithms that use machine learning techniques to analyze and determine any anomalies in the system. The system helps to improve process performance and reduce power consumption by identifying & addressing potential issues at their early stages.
The proceedings contain 59 papers. The special focus in this conference is on HCI in Games. The topics include: Research on the Quantization of User Experience of Spectator Mode in Moba Games;computer-Aided Games-Base...
ISBN:
(纸本)9783030772765
The proceedings contain 59 papers. The special focus in this conference is on HCI in Games. The topics include: Research on the Quantization of User Experience of Spectator Mode in Moba Games;computer-Aided Games-Based Learning for Children with Autism;analysis of the Competitiveness of Asymmetric Games in the Market;a Systematic Review of the Effect of Gamification on Adherence Across Disciplines;an Exploration of the Fear of Attack Strategy in Chess and Its Influence on Class-A Players of Different Chess Personalities: An Exploration Using Virtual Humans;the Factorial Structure and Underlying Contributors of Parents’ Behavioral Involvement in Children’s Video Game Use;in-Game Virtual Consumption and Online Video Game Addiction: A Conceptual Model;Player Types and Game Element Preferences: Investigating the Relationship with the Gamification User Types HEXAD Scale;the Foundations and Frontiers of Research on the Effect of Video Games on Child Development: A Scientometrics and Knowledge-Mapping Analysis Based on CiteSpace;a Specific Measurable Model: How Can Test Results be Influenced by Interactive Prototypes and Design Manuscripts?;using Neural-Network-Driven Image recognition Software to Detect Emotional Reactions in the Face of a Player While Playing a Horror Video Game;exploratory and Confirmatory Factor Analysis of the Chinese Young Children’s Video-Gaming Questionnaire;using Multiple Data Streams in Executive Function Training Games to Optimize Outcomes for Neurodiverse Populations;gameplay as Network: Understanding the Consequences of Automation on Play and Use;hitboxes: A Survey About Collision Detection in Video Games;adaptive Gamification and Its Impact on Performance;analyzing and Prioritizing Usability Issues in Games.
暂无评论