Pub Date : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231682
Bouthayna Msellmi, Daniele Picone, Zouhaier Ben Rabah, M. Mura, I. Farah
In this paper we tackle the problem of data analysis over the higher dimensional space provided by hyperspectral images. Remote sensing data analysis is a complex task due to numerous factors such as the large spectral and spatial diversity. The latter is the key focus of attention in this paper.As a matter of fact, mixed pixels are often sources of uncertainty which affects the accuracy of several approaches whose target is to solve the sub-pixel problem. Although spectral un-mixing techniques can provide abundance fractions within mixed pixels to each class, their associated spatial distribution remains unknown. The set of techniques aimed to solve the above mentioned problem is commonly known as sub-pixel mapping (SPM); existing algorithms based on the spatial dependence assumption cannot solve these problems efficiently and cannot provide a unique configuration for the same problem. In the context of variational framework to solve inverse problems, various strategies were proposed to avoid their intrinsic ill-posedness in the form of regularization. Differently from previous approaches of literature, which apply spatial regularization individually for each class, the proposed method takes also into account spatial links among classes. In order to improve sub-pixel mapping accuracy and, consequently, enhance hyperspectral image classification, we propose a method based on a pre-constructed spectral dictionary and isotropic total variation minimization of classes within and between pixels (SMSD-ITV). Experimental results with real and simulated data sets show the attributes of using spectral dictionary with total variation as a prior model, which lead to improve sub-pixel mapping of different classes tacking into account spatial correlation between them.
{"title":"Sub-pixel Mapping Method based on Total Variation Minimization and Spectral Dictionary","authors":"Bouthayna Msellmi, Daniele Picone, Zouhaier Ben Rabah, M. Mura, I. Farah","doi":"10.1109/ATSIP49331.2020.9231682","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231682","url":null,"abstract":"In this paper we tackle the problem of data analysis over the higher dimensional space provided by hyperspectral images. Remote sensing data analysis is a complex task due to numerous factors such as the large spectral and spatial diversity. The latter is the key focus of attention in this paper.As a matter of fact, mixed pixels are often sources of uncertainty which affects the accuracy of several approaches whose target is to solve the sub-pixel problem. Although spectral un-mixing techniques can provide abundance fractions within mixed pixels to each class, their associated spatial distribution remains unknown. The set of techniques aimed to solve the above mentioned problem is commonly known as sub-pixel mapping (SPM); existing algorithms based on the spatial dependence assumption cannot solve these problems efficiently and cannot provide a unique configuration for the same problem. In the context of variational framework to solve inverse problems, various strategies were proposed to avoid their intrinsic ill-posedness in the form of regularization. Differently from previous approaches of literature, which apply spatial regularization individually for each class, the proposed method takes also into account spatial links among classes. In order to improve sub-pixel mapping accuracy and, consequently, enhance hyperspectral image classification, we propose a method based on a pre-constructed spectral dictionary and isotropic total variation minimization of classes within and between pixels (SMSD-ITV). Experimental results with real and simulated data sets show the attributes of using spectral dictionary with total variation as a prior model, which lead to improve sub-pixel mapping of different classes tacking into account spatial correlation between them.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129819255","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231565
Tan Truong-Ngoc, A. Khenchaf, F. Comblet, Pierre Franck, Jean-Marc Champeyroux, O. Reichert
This paper presents data fusion from multiple Global Navigation Satellite System (GNSS) constellations. GNSS brings more signals and more satellites to improve the accuracy of user’s position. However, multiple failures in satellite’s signals sometimes negatively impact the determination of the user’s position and should be considered. For this purpose, the present paper provides robust Extended Kalman Filter (robust-EKF) to eliminate the outliers. The algorithms are tested by using GPS, Galileo and GLONASS data corresponding on data from base station GRAC in Grasse, France. Applying the robust-EKF method as well as the robust combination of GPS, Galileo, and GLONASS data improves the position accuracy by about 30.0%, 20.7%, and 90% compared to the use of GPS data only, Galileo data only, and GLONASS data only, respectively, and by about 67% compared to the nonrobust combination of GPS, Galileo, and GLONASS data.
{"title":"Robust GPS/Galileo/GLONASS Data Fusion Using Extended Kalman Filter","authors":"Tan Truong-Ngoc, A. Khenchaf, F. Comblet, Pierre Franck, Jean-Marc Champeyroux, O. Reichert","doi":"10.1109/ATSIP49331.2020.9231565","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231565","url":null,"abstract":"This paper presents data fusion from multiple Global Navigation Satellite System (GNSS) constellations. GNSS brings more signals and more satellites to improve the accuracy of user’s position. However, multiple failures in satellite’s signals sometimes negatively impact the determination of the user’s position and should be considered. For this purpose, the present paper provides robust Extended Kalman Filter (robust-EKF) to eliminate the outliers. The algorithms are tested by using GPS, Galileo and GLONASS data corresponding on data from base station GRAC in Grasse, France. Applying the robust-EKF method as well as the robust combination of GPS, Galileo, and GLONASS data improves the position accuracy by about 30.0%, 20.7%, and 90% compared to the use of GPS data only, Galileo data only, and GLONASS data only, respectively, and by about 67% compared to the nonrobust combination of GPS, Galileo, and GLONASS data.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126196780","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231632
Wissal Ben Marzouka, B. Solaiman, A. Hammouda, Zouhour Ben Dhief, K. Bsaïes
This paper aims to present a possibilistic registration method using BRISK. BRISK method is a key point detector and descriptor. It is rotation and scale invariant, but it takes more time to detect the feature points and it suffers from the high number of outliers. The main idea of the proposed method is to apply the theory of possibilities for extracting primitives to obtaining an efficient registration. We explore the suitability of the BRISK method for the task of image registration by limiting the outlier’s number. The proposed method uses the semantic aspect of images for features detection as well as matching. This “semantic focussing process” allows reducing the quantity of information, as well as the noise effects during the matching process by the creation of a new space called “Semantic knowledge space” which contains a set of projections of images each presenting a single content called a “possibilistic maps The experiments as well as the comparative study carried out, using medical images, show the efficiency of the proposed method in terms of outliers’ reduction, noise robustness, time complexity and precision improved.
{"title":"Possibilistic BRISK method for an efficient registration (PBRISK)","authors":"Wissal Ben Marzouka, B. Solaiman, A. Hammouda, Zouhour Ben Dhief, K. Bsaïes","doi":"10.1109/ATSIP49331.2020.9231632","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231632","url":null,"abstract":"This paper aims to present a possibilistic registration method using BRISK. BRISK method is a key point detector and descriptor. It is rotation and scale invariant, but it takes more time to detect the feature points and it suffers from the high number of outliers. The main idea of the proposed method is to apply the theory of possibilities for extracting primitives to obtaining an efficient registration. We explore the suitability of the BRISK method for the task of image registration by limiting the outlier’s number. The proposed method uses the semantic aspect of images for features detection as well as matching. This “semantic focussing process” allows reducing the quantity of information, as well as the noise effects during the matching process by the creation of a new space called “Semantic knowledge space” which contains a set of projections of images each presenting a single content called a “possibilistic maps The experiments as well as the comparative study carried out, using medical images, show the efficiency of the proposed method in terms of outliers’ reduction, noise robustness, time complexity and precision improved.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128177021","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231709
Wafa Baccouch, S. Oueslati, S. Labidi, Bassel Solaiman
Detecting and quantifying abnormalities in the movement of the heart walls such as hypokinesia, akinesia and dyskinesia and measuring their severity is a critical step in the assessment and treatment of ischemic and non-ischemic heart disease. These so-called contraction abnormalities are generally manifested by a decrease in the amplitude of the cardiac contraction reflecting hypokinesia and a complete absence of wall movement indicating akinesia. In case of dyskinesia, the wall is characterized by an abnormal movement, most often ventricular. In the non-pathological case, when the ventricle contracts in systole, it thickens and tends to approach the center of the cavity while in case of dyskinesia it tends to move away. In medical imaging, several methods for regional assessment of cardiac contractile function have been developed. The aim of this article is to review the most relevant approaches available in magnetic resonance imaging (MRI) such as parametric imaging, cardiac contour segmentation and deep learning. At the end of this study, we compared the previously mentioned approaches after explaining their principles, their advantages and disadvantages. The comparison showed that deep learning represents the most precise method in terms of segmentation and quantification of the contraction anomalies.
{"title":"Quantification of left ventricular function in MRI: a review of current approaches","authors":"Wafa Baccouch, S. Oueslati, S. Labidi, Bassel Solaiman","doi":"10.1109/ATSIP49331.2020.9231709","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231709","url":null,"abstract":"Detecting and quantifying abnormalities in the movement of the heart walls such as hypokinesia, akinesia and dyskinesia and measuring their severity is a critical step in the assessment and treatment of ischemic and non-ischemic heart disease. These so-called contraction abnormalities are generally manifested by a decrease in the amplitude of the cardiac contraction reflecting hypokinesia and a complete absence of wall movement indicating akinesia. In case of dyskinesia, the wall is characterized by an abnormal movement, most often ventricular. In the non-pathological case, when the ventricle contracts in systole, it thickens and tends to approach the center of the cavity while in case of dyskinesia it tends to move away. In medical imaging, several methods for regional assessment of cardiac contractile function have been developed. The aim of this article is to review the most relevant approaches available in magnetic resonance imaging (MRI) such as parametric imaging, cardiac contour segmentation and deep learning. At the end of this study, we compared the previously mentioned approaches after explaining their principles, their advantages and disadvantages. The comparison showed that deep learning represents the most precise method in terms of segmentation and quantification of the contraction anomalies.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127453659","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231588
Abir Chaari, Khouloud Kammoun, I.F. Kallel, M. Frikha, S. Kammoun, J. Feki
The visual activity contributes to the perception, identification and localization of the different scenes of life. Any visual impairment can be very bothersome in the daily life and can lead in some cases to a danger like keratoconus.OCT presents a better way of diagnosing of this disease. It is a powerful aid tool for ophthalmic physicians in early decisionmaking for better patient management.In this paper, we used many filters such as Lee, kuan, Frost filters, the Hard thresholding, Soft thresholding and Anisotropic diffusion methods. These filters have the combined property of edge preservation and noise removal.then we present an automatic method of segmentation for the detection of keratoconus
{"title":"Performance Evaluation of Various Denoising Filters and segmentation methods for OCT images","authors":"Abir Chaari, Khouloud Kammoun, I.F. Kallel, M. Frikha, S. Kammoun, J. Feki","doi":"10.1109/ATSIP49331.2020.9231588","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231588","url":null,"abstract":"The visual activity contributes to the perception, identification and localization of the different scenes of life. Any visual impairment can be very bothersome in the daily life and can lead in some cases to a danger like keratoconus.OCT presents a better way of diagnosing of this disease. It is a powerful aid tool for ophthalmic physicians in early decisionmaking for better patient management.In this paper, we used many filters such as Lee, kuan, Frost filters, the Hard thresholding, Soft thresholding and Anisotropic diffusion methods. These filters have the combined property of edge preservation and noise removal.then we present an automatic method of segmentation for the detection of keratoconus","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130369144","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231883
S. Bouzid, Y. Serrestou, K. Raoof, Mohamed Nazih Omri
Wireless sensor nodes are battery-powered devices which makes the design of energy-efficient Wireless Sensor Networks (WSNs) a very challenging issue. In this paper, we propose a new routing protocol for WSN based on distributed Reinforcement Learning (RL). The proposed approach optimises WSN lifetime and energy consumption. This routing protocol learns, over time, the optimal path to the sink node(s). With a dynamic path selection, our algorithm ensures higher energy efficiency, postpones nodes death and isolation. We consider while routing messages the distance between nodes, available energy and hop count to the sink node. The effectiveness of the proposed protocol is demonstrated through simulations and comparisons with some existing algorithms over different lifetime definitions.
{"title":"Efficient Routing Protocol for Wireless Sensor Network based on Reinforcement Learning","authors":"S. Bouzid, Y. Serrestou, K. Raoof, Mohamed Nazih Omri","doi":"10.1109/ATSIP49331.2020.9231883","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231883","url":null,"abstract":"Wireless sensor nodes are battery-powered devices which makes the design of energy-efficient Wireless Sensor Networks (WSNs) a very challenging issue. In this paper, we propose a new routing protocol for WSN based on distributed Reinforcement Learning (RL). The proposed approach optimises WSN lifetime and energy consumption. This routing protocol learns, over time, the optimal path to the sink node(s). With a dynamic path selection, our algorithm ensures higher energy efficiency, postpones nodes death and isolation. We consider while routing messages the distance between nodes, available energy and hop count to the sink node. The effectiveness of the proposed protocol is demonstrated through simulations and comparisons with some existing algorithms over different lifetime definitions.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"216 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133545467","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231698
M. Tlig, M. Bouchouicha, M. Sayadi, E. Moreau
The purpose of this paper is to develop a tracking system. The designed platform is based on two kinds of physical sensors: a doppler radar module and HD Camera. All the measurements from those modules are processed using a data fusion method. In this work, a method based on the Gaussian mixture model is used for foreground detection (i.e. background subtraction). After that our move to a filtering step which is used to refine the detection results firstly obtained, then a tracking process is introduced. As a final stage, all the vision-based measurements are combined with the processed radar raw data. Here, the goal is to perfectly estimate the target velocity in the real time. Added to target 2D positions, this speed information is considered as a third dimension. This is very useful in many applications such as traffic control, robotics, autonomous vehicles etc. In this work a set of experiments is conducted in order to validate the developed tracking method.
{"title":"Multi-Object tracking based on Kalman Filtering Combining Radar and Image Measurements","authors":"M. Tlig, M. Bouchouicha, M. Sayadi, E. Moreau","doi":"10.1109/ATSIP49331.2020.9231698","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231698","url":null,"abstract":"The purpose of this paper is to develop a tracking system. The designed platform is based on two kinds of physical sensors: a doppler radar module and HD Camera. All the measurements from those modules are processed using a data fusion method. In this work, a method based on the Gaussian mixture model is used for foreground detection (i.e. background subtraction). After that our move to a filtering step which is used to refine the detection results firstly obtained, then a tracking process is introduced. As a final stage, all the vision-based measurements are combined with the processed radar raw data. Here, the goal is to perfectly estimate the target velocity in the real time. Added to target 2D positions, this speed information is considered as a third dimension. This is very useful in many applications such as traffic control, robotics, autonomous vehicles etc. In this work a set of experiments is conducted in order to validate the developed tracking method.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133553626","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231764
Bilel Ameur, M. Belahcene, Sabeur Masmoudi, A. Hamida
Unconstrained Face Verification is still an important problem worth researching. The major challenges such as illumination, pose, occlusion and expression can produce more complex variations in both shape and texture of the face. In this paper, we propose a method based on Monogenic Binary Pattern and Convolutional Neural Network (MBP-CNN) to improve the performance of face recognition system. For each facial image, the proposed method firstly extracts local features using Monogenic Binary Pattern (MBP) which is an excellent and powerful local descriptor compared to the well-recognized Gabor filtering-based LBP models. Then, we use Convolutional Neural Networks which is one of the best representative network architectures of deep learning in the literature, in order to extract more deep features. Thus, the developed MBP-CNN has robustness to variations of illumination, occlusion, pose, expression, texture and shape by combining Monogenic Binary Pattern and convolutional neural network. Moreover, MBP-CNN was more accurately represented by combining global and local information of facial images. Experiments demonstrate that our method provided competitive performance on the LFW database, compared to the others described in the state-of-the-art.
{"title":"Unconstrained Face Verification Based on Monogenic Binary Pattern and Convolutional Neural Network","authors":"Bilel Ameur, M. Belahcene, Sabeur Masmoudi, A. Hamida","doi":"10.1109/ATSIP49331.2020.9231764","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231764","url":null,"abstract":"Unconstrained Face Verification is still an important problem worth researching. The major challenges such as illumination, pose, occlusion and expression can produce more complex variations in both shape and texture of the face. In this paper, we propose a method based on Monogenic Binary Pattern and Convolutional Neural Network (MBP-CNN) to improve the performance of face recognition system. For each facial image, the proposed method firstly extracts local features using Monogenic Binary Pattern (MBP) which is an excellent and powerful local descriptor compared to the well-recognized Gabor filtering-based LBP models. Then, we use Convolutional Neural Networks which is one of the best representative network architectures of deep learning in the literature, in order to extract more deep features. Thus, the developed MBP-CNN has robustness to variations of illumination, occlusion, pose, expression, texture and shape by combining Monogenic Binary Pattern and convolutional neural network. Moreover, MBP-CNN was more accurately represented by combining global and local information of facial images. Experiments demonstrate that our method provided competitive performance on the LFW database, compared to the others described in the state-of-the-art.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133554199","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231795
Nihel Zaaboub, A. Douik
The diabetic retinopathy is one of the most frequent causes of visual damage and vision loss. It can cause blindness in the absence of the diagnosis and the treatment. The automatic detection of the hard exudate in color fundus retinal images is an important task to early diagnosis the diabetic retinopathy. In this paper, a hard exudate detection algorithm is proposed. It is based on the application of a learning method to retinal image with removed optic disk. This paper proposes the use of Random Forest algorithm with a specific parameter from which a binary mask of exudate is obtained after intensity thresholding. It achieves 91.40% for sensitivity and 94.38% for the accuracy.
{"title":"Early Diagnosis of Diabetic Retinopathy using Random Forest Algorithm","authors":"Nihel Zaaboub, A. Douik","doi":"10.1109/ATSIP49331.2020.9231795","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231795","url":null,"abstract":"The diabetic retinopathy is one of the most frequent causes of visual damage and vision loss. It can cause blindness in the absence of the diagnosis and the treatment. The automatic detection of the hard exudate in color fundus retinal images is an important task to early diagnosis the diabetic retinopathy. In this paper, a hard exudate detection algorithm is proposed. It is based on the application of a learning method to retinal image with removed optic disk. This paper proposes the use of Random Forest algorithm with a specific parameter from which a binary mask of exudate is obtained after intensity thresholding. It achieves 91.40% for sensitivity and 94.38% for the accuracy.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133746180","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 : 2020-09-01DOI: 10.1109/ATSIP49331.2020.9231677
A. Atitallah, Manel Kammoun
In recent years, most of developers work continuously in order to improve design performances in term of energy efficiency and performances. Therefore, the adoption of high-level synthesis (HLS) techniques can help to reach these requirements, especially when dealing with such complex applications like High Efficiency Video Coding standard (HEVC).This paper discusses a case study of intra prediction algorithm of HEVC using HLS designing method. For this experiment, we used the version 10 of HEVC Test Model (HM) reference software which involves more than 300 functions and over 9000 lines of code. Moreover, this algorithm is implemented in SW/HW environment using Xilinx ZC 702 based-platform. Finally, the experimental results prove that the hardware implementation is able to process 51 video frames per seconde of Full HD (1920xl080p) resolution. However, the SW/HW design can only decode 15 frames per second for 240p video resolutions with a gain of 6% in frame rate and 70% in power consumption relative to SW implementation.
{"title":"High-level design of HEVC intra prediction algorithm","authors":"A. Atitallah, Manel Kammoun","doi":"10.1109/ATSIP49331.2020.9231677","DOIUrl":"https://doi.org/10.1109/ATSIP49331.2020.9231677","url":null,"abstract":"In recent years, most of developers work continuously in order to improve design performances in term of energy efficiency and performances. Therefore, the adoption of high-level synthesis (HLS) techniques can help to reach these requirements, especially when dealing with such complex applications like High Efficiency Video Coding standard (HEVC).This paper discusses a case study of intra prediction algorithm of HEVC using HLS designing method. For this experiment, we used the version 10 of HEVC Test Model (HM) reference software which involves more than 300 functions and over 9000 lines of code. Moreover, this algorithm is implemented in SW/HW environment using Xilinx ZC 702 based-platform. Finally, the experimental results prove that the hardware implementation is able to process 51 video frames per seconde of Full HD (1920xl080p) resolution. However, the SW/HW design can only decode 15 frames per second for 240p video resolutions with a gain of 6% in frame rate and 70% in power consumption relative to SW implementation.","PeriodicalId":384018,"journal":{"name":"2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134448786","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：