Pub Date : 2018-09-01DOI: 10.23919/SPA.2018.8563400
G. Sarwas, S. Skoneczny
In this paper a method for clustering face images based on fractional order SIFT algorithm (FSIFT) is presented. This new approach is based on the dissimilarity matrix. This matrix is constructed by using descriptors calculated for keypoints detected by FSIFT algorithm using derivatives of non integer order. To proof and compared the quality of achieved results the relative error ratio and the F-measure were applying. The final scores of experiments were compared with hierarchical clustering methods based on SIFT and SURF detectors.
{"title":"FSIFT based feature points for face hierarchical clustering","authors":"G. Sarwas, S. Skoneczny","doi":"10.23919/SPA.2018.8563400","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563400","url":null,"abstract":"In this paper a method for clustering face images based on fractional order SIFT algorithm (FSIFT) is presented. This new approach is based on the dissimilarity matrix. This matrix is constructed by using descriptors calculated for keypoints detected by FSIFT algorithm using derivatives of non integer order. To proof and compared the quality of achieved results the relative error ratio and the F-measure were applying. The final scores of experiments were compared with hierarchical clustering methods based on SIFT and SURF detectors.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126616689","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563380
Marcin Grochowina, L. Leniowska
The article presents an innovative solution enabling automatic assessment of the condition of arteriovenous fistula on the basis of the sound emitted by blood flowing inside the fistula. Software and hardware tools included in the designed and manufactured prototype of the device were discussed. The methods currently used in medical practice, based on ultrasound imaging techniques, require the use of specialised equipment supported by medical staff. Routine auscultation with the use of a stethoscope is a method burdened with the subjectivism of a doctor and does not allow to compare the results with historical research. The advantage of the presented solution is the possibility of obtaining a quick, reliable and objective diagnosis. The examination may be carried out by unqualified personnel and may allow the collection of results of historical studies in order to follow the trend of changes in the fistula status.
{"title":"Design and implementation of a device supporting automatic diagnosis of arteriovenous fistula","authors":"Marcin Grochowina, L. Leniowska","doi":"10.23919/SPA.2018.8563380","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563380","url":null,"abstract":"The article presents an innovative solution enabling automatic assessment of the condition of arteriovenous fistula on the basis of the sound emitted by blood flowing inside the fistula. Software and hardware tools included in the designed and manufactured prototype of the device were discussed. The methods currently used in medical practice, based on ultrasound imaging techniques, require the use of specialised equipment supported by medical staff. Routine auscultation with the use of a stethoscope is a method burdened with the subjectivism of a doctor and does not allow to compare the results with historical research. The advantage of the presented solution is the possibility of obtaining a quick, reliable and objective diagnosis. The examination may be carried out by unqualified personnel and may allow the collection of results of historical studies in order to follow the trend of changes in the fistula status.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130334549","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563382
Michal Luczynski, S. Brachmański, A. Dobrucki
In this paper algorithms for eliminating tonal components from acoustic signals are presented. The tonal components are periodic signals with randomly modulated amplitude, phase and frequency. Active elimination of a single tonal component bases on adding a synthesized component with opposite polarity. To do this effectively, it is necessary to meet the criteria for adjusting the values of the parameters for both components. Otherwise, the level reduction may be insufficient or even the level of the component may increase. It is important to define acceptable error values for which are met the reduction criteria. The problem defined in this way allows to indicate the limitations of the method and to create a mathematical model allowing to predict effectiveness for any case in the scope of the analyzed events. As part of the experiment, signals are analyzed with one or more tonal components. The obtained results of the applied algorithm are promising and show the desirability of continuing the work. In special conditions, the developed algorithm allows for reduction of tonal components to values below the acoustic background.
{"title":"Active elimination of tonal components in acoustic signals","authors":"Michal Luczynski, S. Brachmański, A. Dobrucki","doi":"10.23919/SPA.2018.8563382","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563382","url":null,"abstract":"In this paper algorithms for eliminating tonal components from acoustic signals are presented. The tonal components are periodic signals with randomly modulated amplitude, phase and frequency. Active elimination of a single tonal component bases on adding a synthesized component with opposite polarity. To do this effectively, it is necessary to meet the criteria for adjusting the values of the parameters for both components. Otherwise, the level reduction may be insufficient or even the level of the component may increase. It is important to define acceptable error values for which are met the reduction criteria. The problem defined in this way allows to indicate the limitations of the method and to create a mathematical model allowing to predict effectiveness for any case in the scope of the analyzed events. As part of the experiment, signals are analyzed with one or more tonal components. The obtained results of the applied algorithm are promising and show the desirability of continuing the work. In special conditions, the developed algorithm allows for reduction of tonal components to values below the acoustic background.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114616141","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563379
A. Kim, O. Lukovenkova, Y. Marapulets, A. Tristanov
Impulse nature signals generated by complex geophysical systems require special methods to study their internal structure. These signals are characterized by a short duration of impulses and the variability of their structure. The use of classical spectral and time-frequency methods raises great difficulty. The authors propose a model of an impulse signal based on a sparse approximation and an algorithm for identifying a model. The algorithm is a modified matching pursuit algorithm using a physically based system of functions (dictionary). The study of modeling results consists in estimating the time-frequency characteristics of the model components. The paper gives an example of the model application on geoacoustic emission signals of a seismically active region (Kamchatka peninsula). The proposed model and approaches to the model investigation can be used for a wide range of impulse nature signals.
{"title":"Using a sparse model to evaluate the internal structure of impulse signals","authors":"A. Kim, O. Lukovenkova, Y. Marapulets, A. Tristanov","doi":"10.23919/SPA.2018.8563379","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563379","url":null,"abstract":"Impulse nature signals generated by complex geophysical systems require special methods to study their internal structure. These signals are characterized by a short duration of impulses and the variability of their structure. The use of classical spectral and time-frequency methods raises great difficulty. The authors propose a model of an impulse signal based on a sparse approximation and an algorithm for identifying a model. The algorithm is a modified matching pursuit algorithm using a physically based system of functions (dictionary). The study of modeling results consists in estimating the time-frequency characteristics of the model components. The paper gives an example of the model application on geoacoustic emission signals of a seismically active region (Kamchatka peninsula). The proposed model and approaches to the model investigation can be used for a wide range of impulse nature signals.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127927285","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563364
Tomasz Grzywalski, S. Drgas
In this paper a recurrent U-net neural architecture is proposed to speech enhancement. The mentioned neural network architecture is trained to provide a mapping between a spectrogram of a noisy speech and both spectrograms of isolated speech and noise. Some key design choices are being evaluated in experiments and discussed, including: number of levels of the U-net, presence/absence of recurrent layers, presence/absence of max pooling layers as well and upsampling algorithm used in decoder part of the network.
{"title":"Application of recurrent U-net architecture to speech enhancement","authors":"Tomasz Grzywalski, S. Drgas","doi":"10.23919/SPA.2018.8563364","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563364","url":null,"abstract":"In this paper a recurrent U-net neural architecture is proposed to speech enhancement. The mentioned neural network architecture is trained to provide a mapping between a spectrogram of a noisy speech and both spectrograms of isolated speech and noise. Some key design choices are being evaluated in experiments and discussed, including: number of levels of the U-net, presence/absence of recurrent layers, presence/absence of max pooling layers as well and upsampling algorithm used in decoder part of the network.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132347268","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563384
K. Sozański, A. Sozańska
This paper discusses the problem of low frequency measurements of loudspeakers. This can be achieved by performing outdoor measurements (in open-air), by using a near-field technique or by using a large anechoic acoustic chamber. The frequency characteristics of an exemplary loudspeaker are obtained and compared using these three methods. This article also describes anechoic acoustic chamber built in the Science and Technology Park of the University of Zielona Gora.
{"title":"Low Frequency Loudspeaker Measurements Using An Anechoic Acoustic Chamber","authors":"K. Sozański, A. Sozańska","doi":"10.23919/SPA.2018.8563384","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563384","url":null,"abstract":"This paper discusses the problem of low frequency measurements of loudspeakers. This can be achieved by performing outdoor measurements (in open-air), by using a near-field technique or by using a large anechoic acoustic chamber. The frequency characteristics of an exemplary loudspeaker are obtained and compared using these three methods. This article also describes anechoic acoustic chamber built in the Science and Technology Park of the University of Zielona Gora.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133174228","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563404
P. Janus, T. Kryjak
In this paper a hardware implementation of the Gaussian Mixture Model algorithm for background modelling and foreground object segmentation is presented. The proposed vision system is able to handle video stream with resolution up to 4K ($3840 times 2160$ pixels) and 60 frames per second. Moreover, the constraints caused by memory bandwidth limit are also discussed and a few different solutions to tackle this issue have been considered. The designed modules have been verified on the ZCU 102 development board with Xilinx Zynq UltraScale+ MPSoC device. Additionally, the computing performance and power consumption have been estimated.
{"title":"Hardware implementation of the Gaussian Mixture Model foreground object segmentation algorithm working with ultra-high resolution video stream in real-time","authors":"P. Janus, T. Kryjak","doi":"10.23919/SPA.2018.8563404","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563404","url":null,"abstract":"In this paper a hardware implementation of the Gaussian Mixture Model algorithm for background modelling and foreground object segmentation is presented. The proposed vision system is able to handle video stream with resolution up to 4K ($3840 times 2160$ pixels) and 60 frames per second. Moreover, the constraints caused by memory bandwidth limit are also discussed and a few different solutions to tackle this issue have been considered. The designed modules have been verified on the ZCU 102 development board with Xilinx Zynq UltraScale+ MPSoC device. Additionally, the computing performance and power consumption have been estimated.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127342488","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563399
M. Parfieniuk, S. Park
Recently, rough approximations of the Discrete Cosine Transform (DCT) have been proposed that can be implemented as multiplier-less, low-area, and low-power circuits. Promoters of such algorithms considered simpler and simpler data-flow graphs, by using fewer and fewer additions and bit-shifts compared to finer approximations developed at the turn of the 20th and 21th centuries. However, they neglected to carefully check whether an approximation works like the original, and from another point of view, they ignore well-known essential results of the theory and practice of image transforms. This paper shows that one of such solutions is not as perfect as advertised, or even seems to be useless, suffering from inherent disadvantages of non-selective filters and non-smooth basis functions. We point out what is lacking in the published evaluations of the algorithm and analyse its properties, demonstrating that it behaves differently from the DCT and thus is suitable to neither image compression nor pattern recognition. In particular, we show that it poorly decorrelates samples of natural images, and unpleasant in-block artefacts appear in decoded pictures.
{"title":"A critique of some rough approximations of the DCT","authors":"M. Parfieniuk, S. Park","doi":"10.23919/SPA.2018.8563399","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563399","url":null,"abstract":"Recently, rough approximations of the Discrete Cosine Transform (DCT) have been proposed that can be implemented as multiplier-less, low-area, and low-power circuits. Promoters of such algorithms considered simpler and simpler data-flow graphs, by using fewer and fewer additions and bit-shifts compared to finer approximations developed at the turn of the 20th and 21th centuries. However, they neglected to carefully check whether an approximation works like the original, and from another point of view, they ignore well-known essential results of the theory and practice of image transforms. This paper shows that one of such solutions is not as perfect as advertised, or even seems to be useless, suffering from inherent disadvantages of non-selective filters and non-smooth basis functions. We point out what is lacking in the published evaluations of the algorithm and analyse its properties, demonstrating that it behaves differently from the DCT and thus is suitable to neither image compression nor pattern recognition. In particular, we show that it poorly decorrelates samples of natural images, and unpleasant in-block artefacts appear in decoded pictures.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130116675","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563418
Yaprak Eminaga, Adem Coskun, I. Kale
ElectroCardioGram (ECG) signals are widely used for diagnostic purposes. However, it is well known that these recordings are usually corrupted with different type of noise/artifacts which might lead to misdiagnosis of the patient. This paper presents the design and novel use of Infinite Impulse Response (IIR) filter based Discrete Wavelet Transform (DWT) for ECG denoising that can be employed in ambulatory health monitoring applications. The proposed system is evaluated and compared in terms of denoising performance as well as the computational complexity with the conventional Finite Impulse Response (FIR) based DWT systems. For this purpose, raw ECG data from MIT-BIH arrhythmia database are contaminated with synthetic noise and denoised with the aforementioned filter banks. The results from 100 Monte Carlo simulations demonstrated that the proposed filter banks provide better denoising performance with fewer arithmetic operations than those reported in the open literature.
{"title":"IIR Wavelet Filter Banks for ECG Signal Denoising","authors":"Yaprak Eminaga, Adem Coskun, I. Kale","doi":"10.23919/SPA.2018.8563418","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563418","url":null,"abstract":"ElectroCardioGram (ECG) signals are widely used for diagnostic purposes. However, it is well known that these recordings are usually corrupted with different type of noise/artifacts which might lead to misdiagnosis of the patient. This paper presents the design and novel use of Infinite Impulse Response (IIR) filter based Discrete Wavelet Transform (DWT) for ECG denoising that can be employed in ambulatory health monitoring applications. The proposed system is evaluated and compared in terms of denoising performance as well as the computational complexity with the conventional Finite Impulse Response (FIR) based DWT systems. For this purpose, raw ECG data from MIT-BIH arrhythmia database are contaminated with synthetic noise and denoised with the aforementioned filter banks. The results from 100 Monte Carlo simulations demonstrated that the proposed filter banks provide better denoising performance with fewer arithmetic operations than those reported in the open literature.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128826292","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 : 2018-09-01DOI: 10.23919/SPA.2018.8563419
J. Balcerek, Mateusz Luczak, P. Pawlowski, A. Dabrowski
This paper presents a utilization of the stereovision in the automotive applications and video monitoring. Two video channels of stereovision system offer a new quality of recognition of image details. Even with popular object recognition algorithms the stereovision significantly improves the analysis plausibility. Preliminary experiments with the specially prepared 3D video test database show improvement of the recognition effectiveness by 10-30% in comparison to the single video channel.
{"title":"Automatic recognition of image details using stereovision and 2D algorithms","authors":"J. Balcerek, Mateusz Luczak, P. Pawlowski, A. Dabrowski","doi":"10.23919/SPA.2018.8563419","DOIUrl":"https://doi.org/10.23919/SPA.2018.8563419","url":null,"abstract":"This paper presents a utilization of the stereovision in the automotive applications and video monitoring. Two video channels of stereovision system offer a new quality of recognition of image details. Even with popular object recognition algorithms the stereovision significantly improves the analysis plausibility. Preliminary experiments with the specially prepared 3D video test database show improvement of the recognition effectiveness by 10-30% in comparison to the single video channel.","PeriodicalId":265587,"journal":{"name":"2018 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123090793","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：