Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743336
Jose D. Cortes, Yulieth Jimenez, C. Duarte
Nonintrusive Load Monitoring (NILM) provides information about the electrical power consumption per appliance in a house to manage the energy consumption. NILM requires measurements in only one point and algorithms to make load disaggregation. One approach is classifying characteristics of the appliance through machine learning techniques such as Support Vector Machines (SVM) and Artificial Neural Networks (ANN). These techniques have limitations in the database use and the disregard of the information context. In this paper a reasoning technique based on the Case Based Reasoning (CBR) reasoner called HYPO is proposed. This reasoner creates hypothetical cases to classify new cases based on the solution of previous experiences. The study is focused on lighting appliances which represent meaningful power consumption in the houses. Electrical measurements lamps in steady state were acquired in the Laboratory, for individual and combined operation. Additionally, characteristics are computed to build the CBR HYPO models. The performance of CBR HYPO is evaluated and compared to the one of SVM. As a result, CBR HYPO outperforms the SVM for combined operation of lamps, while it fails behind SVM for individual operation.
非侵入式负载监控(NILM)提供有关房屋中每个电器的电力消耗的信息，以管理能源消耗。NILM只需要在一个点上进行测量，并使用算法进行负载分解。一种方法是通过机器学习技术(如支持向量机(SVM)和人工神经网络(ANN))对设备的特征进行分类。这些技术在数据库使用方面有局限性，而且不考虑信息上下文。本文提出了一种基于案例推理(Case based reasoning, CBR)推理器的推理技术——HYPO。该推理器根据以往经验的解决方案创建假设案例，对新案例进行分类。这项研究的重点是照明电器，这代表了房屋中有意义的电力消耗。在实验室获得了稳定状态的电测量灯，用于单独和组合操作。此外，计算特征以建立CBR HYPO模型。对CBR HYPO算法的性能进行了评价，并与SVM算法进行了比较。因此，CBR HYPO在灯具组合运行时优于SVM，而在单个运行时落后于SVM。
{"title":"Reasoner design based on HYPO for classification of lighting loads","authors":"Jose D. Cortes, Yulieth Jimenez, C. Duarte","doi":"10.1109/STSIVA.2016.7743336","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743336","url":null,"abstract":"Nonintrusive Load Monitoring (NILM) provides information about the electrical power consumption per appliance in a house to manage the energy consumption. NILM requires measurements in only one point and algorithms to make load disaggregation. One approach is classifying characteristics of the appliance through machine learning techniques such as Support Vector Machines (SVM) and Artificial Neural Networks (ANN). These techniques have limitations in the database use and the disregard of the information context. In this paper a reasoning technique based on the Case Based Reasoning (CBR) reasoner called HYPO is proposed. This reasoner creates hypothetical cases to classify new cases based on the solution of previous experiences. The study is focused on lighting appliances which represent meaningful power consumption in the houses. Electrical measurements lamps in steady state were acquired in the Laboratory, for individual and combined operation. Additionally, characteristics are computed to build the CBR HYPO models. The performance of CBR HYPO is evaluated and compared to the one of SVM. As a result, CBR HYPO outperforms the SVM for combined operation of lamps, while it fails behind SVM for individual operation.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123520684","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743354
Laura Juliana Vargas Escobar, S. A. Salinas, Luis Alfonso Fiallo Duran
This paper presents a proposal of a measurement system that will permit to quantify two gait parameters: short step and stride, within uncontrolled environments. The short step and the stride are two important parameters in the gait analysis that usually are measured in a specialized laboratory, but this measurement can be done in uncontrolled environments where the patient would be more comfortable. In order to calculate the gait parameters without a complex setup, to use an Inertial Measurement Unit (IMU) is proposed, because it is a small and portable sensor, and provides linear acceleration data. These data are adjusted to offset the gravitational forces, and, in further research, they are going to be used to calculate the short step and stride.
{"title":"Proposal of a short step and stride measurement system for uncontrolled environments","authors":"Laura Juliana Vargas Escobar, S. A. Salinas, Luis Alfonso Fiallo Duran","doi":"10.1109/STSIVA.2016.7743354","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743354","url":null,"abstract":"This paper presents a proposal of a measurement system that will permit to quantify two gait parameters: short step and stride, within uncontrolled environments. The short step and the stride are two important parameters in the gait analysis that usually are measured in a specialized laboratory, but this measurement can be done in uncontrolled environments where the patient would be more comfortable. In order to calculate the gait parameters without a complex setup, to use an Inertial Measurement Unit (IMU) is proposed, because it is a small and portable sensor, and provides linear acceleration data. These data are adjusted to offset the gravitational forces, and, in further research, they are going to be used to calculate the short step and stride.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"142 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131561161","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743326
Raul Vargas, J. Pineda, A. Marrugo, L. Romero
In Fourier Transform Profilometry, a filtering procedure is performed to separate the desired information (first order spectrum) from other unwanted contributions such as the background component (zero-order spectrum). However, if the zero-order spectrum and the high order spectra component interfere the fundamental spectra, the 3D reconstruction precision decreases. In this paper, we test two recently proposed methods for removing the background intensity so as to improve Fourier Transform Profilometry reconstruction precision. The first method is based on the twice piece-wise Hilbert transform. The second is based on Bidimensional Empirical Mode Decomposition, but the decomposition is carried out by morphological operations In this work, we present as a novel contribution, the sequential combination of these two methods for removing the background intensity and other unwanted frequencies close to the first order spectrum, thus obtaining the 3D topography of the object. Encouraging experimental results show the advantage of the proposed method.
{"title":"Background intensity removal in structured light three-dimensional reconstruction","authors":"Raul Vargas, J. Pineda, A. Marrugo, L. Romero","doi":"10.1109/STSIVA.2016.7743326","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743326","url":null,"abstract":"In Fourier Transform Profilometry, a filtering procedure is performed to separate the desired information (first order spectrum) from other unwanted contributions such as the background component (zero-order spectrum). However, if the zero-order spectrum and the high order spectra component interfere the fundamental spectra, the 3D reconstruction precision decreases. In this paper, we test two recently proposed methods for removing the background intensity so as to improve Fourier Transform Profilometry reconstruction precision. The first method is based on the twice piece-wise Hilbert transform. The second is based on Bidimensional Empirical Mode Decomposition, but the decomposition is carried out by morphological operations In this work, we present as a novel contribution, the sequential combination of these two methods for removing the background intensity and other unwanted frequencies close to the first order spectrum, thus obtaining the 3D topography of the object. Encouraging experimental results show the advantage of the proposed method.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129170216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743315
Laura Victoria Vigoya Morales, María Dolores Valdés, C. Trillo
Enterococci are part of the normal intestinal flora of humans and animals. They have long been recognized as important human pathogens and are increasing. Enterococcus faecalis and Enterococcus faecium are the most prevalent species cultured from humans, accounting for more than 90 % of clinical isolates. Due to their ubiquity in human feces and persistence in the environment, enterococci have been adopted as indicators of human fecal pollution in water. One of the methods used in water quality control is the membrane filtration technique (Membrane Filtration - MF) (ISO7899-2). This method requires the cultivation of bacteria (enterococci), which is a great disadvantage because the time required to obtain the final result is between 24 and 48 hours. This work proposes a design of a system that detects, with optical sensors, the presence of simulated bacterial colonies in the early stages of the cultures (14-24 h). An image processing system (ZooMat) has been developed with Matlab to detect simulated colonies at early stages, which allows you to process the image before counting. To obtain detection and a count of bacterial colonies on each image, we integrate NICE (an open source, free software) to our system, to gather the results. The entire system allows detection of particles at about 60 μm.
{"title":"Early detection method of enterococci for water quality control with digital image processing techniques","authors":"Laura Victoria Vigoya Morales, María Dolores Valdés, C. Trillo","doi":"10.1109/STSIVA.2016.7743315","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743315","url":null,"abstract":"Enterococci are part of the normal intestinal flora of humans and animals. They have long been recognized as important human pathogens and are increasing. Enterococcus faecalis and Enterococcus faecium are the most prevalent species cultured from humans, accounting for more than 90 % of clinical isolates. Due to their ubiquity in human feces and persistence in the environment, enterococci have been adopted as indicators of human fecal pollution in water. One of the methods used in water quality control is the membrane filtration technique (Membrane Filtration - MF) (ISO7899-2). This method requires the cultivation of bacteria (enterococci), which is a great disadvantage because the time required to obtain the final result is between 24 and 48 hours. This work proposes a design of a system that detects, with optical sensors, the presence of simulated bacterial colonies in the early stages of the cultures (14-24 h). An image processing system (ZooMat) has been developed with Matlab to detect simulated colonies at early stages, which allows you to process the image before counting. To obtain detection and a count of bacterial colonies on each image, we integrate NICE (an open source, free software) to our system, to gather the results. The entire system allows detection of particles at about 60 μm.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"7 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132192175","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743301
Julian Camilo Araque Gomez, Miguel Marquez Castellanos, Henry Arguello Fuentes
Computed tomography (CT) is a non-destructive technique that allows estimation and visualization of the internal structure of an object. Traditionally, CT images are captured by a CT scanner. However, different factors reduce the quality of the acquired images. To obtain a high quality CT images is necessary increase the number of sensors or oversample the object. The number of projections needed for sensing a CT scene is determined by the Nyquist limit, however, in some cases the imposed projections number is excessive. Coded aperture are elements that can block or allow the passing of X-rays and is one approach that can overcome these limitations. Compressive sensing (CS) has emerged as a sampling technique requiring fewer projections than those specified by the Nyquist criterion. CS is a theory to acquire and to reconstruct a signal efficiently by the search of a sparse solution to an indeterminate system of linear equations. A strategy to introduce CS theory in a CT configuration is to include elements into the system that allow coding the measurements to get compressed samples. This paper describes a CS system for CT based on coded apertures using two sources and a two-dimensional array that rotate around the object. An optimized value of transmittance and an aperture distribution are selected such that the quality of reconstruction is efficient. In order to compare the performance of the proposed method, two real CT images and two synthetic CT image were used. Simulations indicate that CT architecture provides comparable results to those achieved with traditional CT architectures. The simulation results show that the proposed method allows more diversity coding. This allows up to 2 dB improvement in terms of PSNR than the results obtained using traditional architecture cone beam.
{"title":"Desing of a coded aperture base computed tomography architecture with two X-ray rotating sources","authors":"Julian Camilo Araque Gomez, Miguel Marquez Castellanos, Henry Arguello Fuentes","doi":"10.1109/STSIVA.2016.7743301","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743301","url":null,"abstract":"Computed tomography (CT) is a non-destructive technique that allows estimation and visualization of the internal structure of an object. Traditionally, CT images are captured by a CT scanner. However, different factors reduce the quality of the acquired images. To obtain a high quality CT images is necessary increase the number of sensors or oversample the object. The number of projections needed for sensing a CT scene is determined by the Nyquist limit, however, in some cases the imposed projections number is excessive. Coded aperture are elements that can block or allow the passing of X-rays and is one approach that can overcome these limitations. Compressive sensing (CS) has emerged as a sampling technique requiring fewer projections than those specified by the Nyquist criterion. CS is a theory to acquire and to reconstruct a signal efficiently by the search of a sparse solution to an indeterminate system of linear equations. A strategy to introduce CS theory in a CT configuration is to include elements into the system that allow coding the measurements to get compressed samples. This paper describes a CS system for CT based on coded apertures using two sources and a two-dimensional array that rotate around the object. An optimized value of transmittance and an aperture distribution are selected such that the quality of reconstruction is efficient. In order to compare the performance of the proposed method, two real CT images and two synthetic CT image were used. Simulations indicate that CT architecture provides comparable results to those achieved with traditional CT architectures. The simulation results show that the proposed method allows more diversity coding. This allows up to 2 dB improvement in terms of PSNR than the results obtained using traditional architecture cone beam.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129846680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743331
Diego Vergara, S. Hernández, Felipe Jorquera
Existing implementations of face recognition systems are created under controlled environments and tested using a limited amount of data. Also, these techniques have a high computational cost which forbids incremental learning that is required in real-time. We propose a gender estimation implementation based on Multinomial Naive Bayes and Local Binary Patterns. The method is tested in a modern age and gender recognition dataset with realistic examples. In order to get state-of-the-art results, Adaboost is also proposed and tested.
{"title":"Multinomial Naive Bayes for real-time gender recognition","authors":"Diego Vergara, S. Hernández, Felipe Jorquera","doi":"10.1109/STSIVA.2016.7743331","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743331","url":null,"abstract":"Existing implementations of face recognition systems are created under controlled environments and tested using a limited amount of data. Also, these techniques have a high computational cost which forbids incremental learning that is required in real-time. We propose a gender estimation implementation based on Multinomial Naive Bayes and Local Binary Patterns. The method is tested in a modern age and gender recognition dataset with realistic examples. In order to get state-of-the-art results, Adaboost is also proposed and tested.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124436095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/STSIVA.2016.7743350
Santiago González, T. R. Vargas, P. Arce, J. C. Guerri
This paper presents the design and implementation of a set of prototype nodes that have the ability to establish communication links in ad hoc mode. The prototypes were implemented using low cost, single board computers with embedded Linux (Raspberry Pi devices). The implemented stations aim to set a wireless sensor network for the capture of variables applied to agricultural environments. In particular, a camera module has been included on a node for remote video monitoring of farming zones, and also a GPS module for the capture of geolocation information. Nodes can be accessed remotely by means of the developed web interface. The routing process between nodes is carried out using the S-OLSR mechanism (OLSR modification) in order to set up routes taking into account the energy limitations as well as the location of each device in the topology. Results describe the contribution of this work to the design of monitoring applications on agricultural zones by means of the deployment of autonomous ad hoc nodes and energy routing optimization.
{"title":"Energy optimization for video monitoring system in agricultural areas using single board computer nodes and wireless ad hoc networks","authors":"Santiago González, T. R. Vargas, P. Arce, J. C. Guerri","doi":"10.1109/STSIVA.2016.7743350","DOIUrl":"https://doi.org/10.1109/STSIVA.2016.7743350","url":null,"abstract":"This paper presents the design and implementation of a set of prototype nodes that have the ability to establish communication links in ad hoc mode. The prototypes were implemented using low cost, single board computers with embedded Linux (Raspberry Pi devices). The implemented stations aim to set a wireless sensor network for the capture of variables applied to agricultural environments. In particular, a camera module has been included on a node for remote video monitoring of farming zones, and also a GPS module for the capture of geolocation information. Nodes can be accessed remotely by means of the developed web interface. The routing process between nodes is carried out using the S-OLSR mechanism (OLSR modification) in order to set up routes taking into account the energy limitations as well as the location of each device in the topology. Results describe the contribution of this work to the design of monitoring applications on agricultural zones by means of the deployment of autonomous ad hoc nodes and energy routing optimization.","PeriodicalId":373420,"journal":{"name":"2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127749641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：