Mathematical imageprocessing by technical vision systems in real-time can be divided into two stages: pre-treatment (filtering, contrast, etc.) and final (alignment, visualization, navigation solution to the problem,...
详细信息
ISBN:
(纸本)9781509067428
Mathematical imageprocessing by technical vision systems in real-time can be divided into two stages: pre-treatment (filtering, contrast, etc.) and final (alignment, visualization, navigation solution to the problem, etc.). These problems can be solved by many famous and specially developed methods with varying degrees of effectiveness. We propose a mathematical model and algorithm of automatic choice the most effective method at each stage of the imageprocessing pipeline in relation to the current situation at its input.
The present paper focuses on a multimodal system based on electrooculography, imageprocessing, and speech recognition for simple and real-time controlling of robotic systems such as manipulators. Electrooculogram was...
详细信息
ISBN:
(纸本)9781538626405
The present paper focuses on a multimodal system based on electrooculography, imageprocessing, and speech recognition for simple and real-time controlling of robotic systems such as manipulators. Electrooculogram was based on voluntary eye movements and processed by thresholds;the average accuracy obtained was about 75 %. Two real-time and simple algorithms are presented for imageprocessing and speech recognition. Color segmentation successfully recognizes red, green, and blue objects. The algorithm used for speech processing represents the correlation of frequency domain for the desired word with the source words. The average accuracy of this algorithm is 90%. It is used for achieving higher degrees of freedom and commanding to the system in order for confirmation and next step recognition. Four different multimodal architectures were designed as combinations of electrooculography, imageprocessing, and speech processingalgorithms. Simulation of the architectures combines human vision (EOG) and robot vision (imageprocessing) in one system to improve accuracy, operation speed, degrees of freedom, orientation efficiency, target recognition, and the accuracy of robot's gripper distance adjustment. At last one of the architectures is implemented on a two degree of-freedom robotic manipulator to evaluate system operation speed and capability. The results indicated that the architectures control the system in a highly efficient manner. It can be thus concluded that a multimodal system based on EOG and imageprocessing improves the accuracy and reliability of rehabilitation and industrial systems with high operation speed.
The proceedings contain 38 papers. The topics discussed include: a system for exploratory analysis in cloud;privacy-preserving and unforgeable searchable encrypted audit logs for cloud storage;a real-time sensor netwo...
ISBN:
(纸本)9781538658505
The proceedings contain 38 papers. The topics discussed include: a system for exploratory analysis in cloud;privacy-preserving and unforgeable searchable encrypted audit logs for cloud storage;a real-time sensor network aggregation computing system;a distributed technique for repairing inconsistencies of functional dependencies;MTComm based virtualization and integration of physical machine operations with digital-twins in cyber-physical manufacturing cloud;difference function projective synchronization for secure communication based on complex chaotic systems;an approach of fog detecting magnitude using referenceless perceptual image defogging;reliability ranking prediction for cloud services via skyline;a cooperative game for online cloud federation formation based on security risk assessment;a knowledge-based approach to reducing complexity of maintaining semantic constraints in data exchange and integration;review on big data fusion methods of quality inspection for consumer goods;unique topic query processing on cloud;enterprise workflow modeling based on priced timed Petri nets;performance evaluation of BATMAN-adv wireless mesh network routing algorithms;FASTBEE: a fast and self-adaptive clustering algorithm towards to edge computing;cryptanalysis of an RFID ownership transfer protocol based on cloud;network security based on d-s evidence theory optimizing CS-BP neural network situation assessment;access control scheme based on combination of blockchain and XOR-coding for ICN;and an effective scheme to detect and prevent tampering on the physical layer of WSN.
The proceedings contain 31 papers. The special focus in this conference is on Web and Big Data. The topics include: Text Classification Methods Based on SVD and FCM;An imageprocessing Method via OpenCL for Identifica...
ISBN:
(纸本)9783030012977
The proceedings contain 31 papers. The special focus in this conference is on Web and Big Data. The topics include: Text Classification Methods Based on SVD and FCM;An imageprocessing Method via OpenCL for Identification of Pulmonary Nodules;Pulmonary Nodule Segmentation Method of CT images Based on 3D-FCN;research on Video Recommendation Algorithm Based on Knowledge Reasoning of Knowledge Graph;a Hybrid Framework for Query processing and Data Analytics on Spark;improving Network-Based Top-N Recommendation with Background Knowledge from Linked Open Data;a Web-Based Theme-Related Word Set Construction Algorithm;classifying Personal Photo Collections: An Event-Based Approach;a Learning Analytic Model for Smart Classroom;a Self-representation Model for Robust Clustering of Categorical Sequences;understanding User Interests Acquisition in Personalized Online Course Recommendation;a Learning Analytics System for Cognition Analysis in Online Learning Community;A Semantic Role Mining and Learning Performance Prediction Method in MOOCs;MOOC Guider: An End-to-End Dialogue System for MOOC Users;AUnet: An Unsupervised Method for Answer Reliability Evaluation in Community QA systems;model and Practice of Crowd-Based Education;Exploring Business Models and Dynamic Pricing Frameworks for SPOC Services;speed-Up algorithms for Happiness-Maximizing Representative Databases;multi-location Influence Maximization in Location-Based Social Networks;emotion Analysis for the Upcoming Response in Open-Domain Human-Computer Conversation;diversified Spatial Keyword Query on Topic Coverage;a Recurrent Neural Network Language Model Based on Word Embedding;inferring Social Ties from Multi-view Spatiotemporal Co-occurrence;sequence-As-Feature Representation for Subspace Classification of Multivariate Time Series;discovering Congestion Propagation Patterns by Co-location Pattern Mining;Spectroscopy-Based Food Internal Quality Evaluation with XGBoost Algorithm.
A key challenge when displaying and processing sensed real-time 3D data is efficiency of generating and post-processingalgorithms in order to acquire high quality 3D content. In contrast, our approach focuses on volu...
详细信息
A key challenge when displaying and processing sensed real-time 3D data is efficiency of generating and post-processingalgorithms in order to acquire high quality 3D content. In contrast, our approach focuses on volumetric generation and processing volumetric data using an efficient low-cost hardware setting. Acquisition of volumetric data is performed by connecting several Kinect v2 scanners to a single PC that are subsequently calibrated using planar pattern. This process is by no means trivial and requires well designed algorithms for fast processing and quick rendering of volumetric data. This can be achieved by fusing efficient filtering methods such as Weighted median filter (WM), Radius outlier removal (ROR) and Laplace-based smoothing algorithm. In this context, we demonstrate the robustness and efficiency of our technique by sensing several scenes.
Denoising using Discrete Wavelet Transform is very popular in imageprocessing. The main tasks of this kind of approach include the image decomposition into random sub-bands and the Wavelet coefficients extraction. In...
详细信息
ISBN:
(纸本)9781538693865;9781538693858
Denoising using Discrete Wavelet Transform is very popular in imageprocessing. The main tasks of this kind of approach include the image decomposition into random sub-bands and the Wavelet coefficients extraction. In this paper a GPU parallel algorithm to compute and extract Wavelet coefficients is proposed. As preprocessing we use the Donoho method to study the amount of noise present in the images; while for suppression or shrinkage criterion we suggest the Garrote threshold. The selection process in this work was performed through two algorithms: Hard and Non-Negative Garrote Thresholding. Our work-frame uses a parallel approach to extract Wavelet coefficients at level of sub-bands, so that significant improvements are achieved over the CPU version, which a speedup gain of around four times with respect to the serial version. Some tests and numerical experiments show the efficiency and the quality obtained using the proposed method.
This paper presents a novel solution for finding a person of interest using a mobile agent platform developed with Raspberry PI iii module in indoor dynamic environments. This method is capable of extracting human fig...
详细信息
ISBN:
(纸本)9781538609699
This paper presents a novel solution for finding a person of interest using a mobile agent platform developed with Raspberry PI iii module in indoor dynamic environments. This method is capable of extracting human figures from complex backgrounds while the platform is still moving and compare the identified human figures against the query image using features like clothing and other accessories. The proposed method processes the video near real time with limited resources on the Raspberry PI iii module and communicate the findings to the central control system using a specially designed communication protocol. System is capable of identifying the person of interest with the accuracy level of more than 90%. The proposed method can either work on its own or be and be integrated with an existing CCTV systems improving its capabilities further.
Super-resolution reconstruction algorithms have been extensively studied for the last years. However, despite the progress made in this field, many issues remain to be solved. Some of them are basically omitted and th...
详细信息
ISBN:
(纸本)9781509063444
Super-resolution reconstruction algorithms have been extensively studied for the last years. However, despite the progress made in this field, many issues remain to be solved. Some of them are basically omitted and their importance is trivialized. Routinely, for instance, researchers are willing to make the relative motion model simpler than it should be considered. The commonly applied non-rigid registration method being manually defined does not capture the real motion characteristics that could occur in image sequences. This work extends Iterative Back Projection (IBP) framework in several ways. It nests image priors, deblurring and a discrete dense displacement sampling for the deformable registration of high-resolution images at its core. Applying these constraints to a global optimum of the cost function can be calculated efficiently exploiting dynamic programming. It leads to the smoothness of the deformations of the image's features. This paper proposes an improved super-resolution method while making no compromise on image quality. The author experiment results confirmed the empirical observations, in particular, that the state of the art registration algorithm and blur and noise estimate procedures, as well as image priors, lead to promising results.
This paper suggests a novel technique for the tool parameter measurement based on machine vision. Tool images are captured by using a machine vision system and the outer contour image of the cutter is obtained by usin...
详细信息
This paper suggests a novel technique for the tool parameter measurement based on machine vision. Tool images are captured by using a machine vision system and the outer contour image of the cutter is obtained by using the machine vision technology. The HALCON imageprocessing library is used as the development platform to build the tool parameters test system. Several algorithms including image segmentation, edge extraction and fitting ellipse determination is used for imageprocessing. The tool parameters such as external diameter and contour angle of tool edge can be obtained after rebuilding the contour of tool edge. The proposed scheme is shown to be reliable and effective for the automated tool parameter measurement.
The basic features of some of the most versatile and popular open source frameworks for machine learning (TensorFlow, Deep Learning4j, and H2O) are considered and compared. Their comparative analysis was performed and...
详细信息
暂无评论