Pub Date : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116893
Amir. M Mousavi. H, A. Bossaghzadeh
One of the machine vision tasks is optical character recognition (OCR) that researchers in this field are trying to achieve a high performance and accuracy in the classification task. In this paper, we have used a fine tuned deep Neural networks for Hoda dataset, which is the largest dataset for Persian handwritten digit classification, to extract valuable discriminative features. then, these features are fed to a linear support vector machine (SVM) for classification part. In the next experiment, In order to improve the accuracy and computational load, we applied the Principal component analysis (PCA) to reduce the extracted features dimensions then we fed it to SVM. To the best of our knowledge the proposed method was better than other methods in terms of accuracy measure
{"title":"Improving Persian Digit Recognition by Combining Deep Neural Networks and SVM and Using PCA","authors":"Amir. M Mousavi. H, A. Bossaghzadeh","doi":"10.1109/MVIP49855.2020.9116893","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116893","url":null,"abstract":"One of the machine vision tasks is optical character recognition (OCR) that researchers in this field are trying to achieve a high performance and accuracy in the classification task. In this paper, we have used a fine tuned deep Neural networks for Hoda dataset, which is the largest dataset for Persian handwritten digit classification, to extract valuable discriminative features. then, these features are fed to a linear support vector machine (SVM) for classification part. In the next experiment, In order to improve the accuracy and computational load, we applied the Principal component analysis (PCA) to reduce the extracted features dimensions then we fed it to SVM. To the best of our knowledge the proposed method was better than other methods in terms of accuracy measure","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"62 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132123540","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116883
Maryam Moradifar, A. Shahbahrami
Denoising is an important process before applying other post-processing techniques on medical images. To obtain better quality images many denoising approaches have been introduced. Gaussian filter is a spatial domain filter, which is proper to blur and to remove noise from images. Since the Gaussian filter modifies the input signal by convolution with a Gaussian function it is a computationally intensive algorithm. Hence to enhance the performance of the algorithm, it is better to perform two 1-D convolution operations instead of one 2-D convolution operation and then parallelize it. In this paper in order to increase the performance of 1-D convolution operation, we exploit both Data- and Thread-Level Parallelism using parallel programming models such as Intrinsic Programming Model, Compiler's Automatic Vectorization and Open Multi-Processing. The experimental results were shown that the performance of our implementations is much higher than other approaches Performance improvements of Multi-threaded version of all implementations are significantly improved compared to single-core implementations, and a speedup of 52.33x obtained over the optimal scalar implementation.
{"title":"Performance Improvement of Gaussian Filter using SIMD Technology","authors":"Maryam Moradifar, A. Shahbahrami","doi":"10.1109/MVIP49855.2020.9116883","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116883","url":null,"abstract":"Denoising is an important process before applying other post-processing techniques on medical images. To obtain better quality images many denoising approaches have been introduced. Gaussian filter is a spatial domain filter, which is proper to blur and to remove noise from images. Since the Gaussian filter modifies the input signal by convolution with a Gaussian function it is a computationally intensive algorithm. Hence to enhance the performance of the algorithm, it is better to perform two 1-D convolution operations instead of one 2-D convolution operation and then parallelize it. In this paper in order to increase the performance of 1-D convolution operation, we exploit both Data- and Thread-Level Parallelism using parallel programming models such as Intrinsic Programming Model, Compiler's Automatic Vectorization and Open Multi-Processing. The experimental results were shown that the performance of our implementations is much higher than other approaches Performance improvements of Multi-threaded version of all implementations are significantly improved compared to single-core implementations, and a speedup of 52.33x obtained over the optimal scalar implementation.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115932878","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116917
Bahman Rouhani, Alireza Samadzadeh, M. Rahmati, A. Nickabadi
Advances in deep learning has lead to drastic improvements in face recognition. A key part of these deep models is their loss function. Consequently developing an efficient and suitable loss function has been an important topic in face recognition in the recent years. Angular-margin-based losses achieve an acceptable performance and inter-class separability. However they are held back by their enforcement of hard margins on all the samples of the training dataset, regardless of whether these samples actually differ from all the other classes enough to enforce a margin. It can be argued that in a large enough dataset with many different settings and age gaps, some faces will look similar to the faces of other classes. In an intuitive and expressive embedding, we expect some faces to be embedded near similar classes with a small margin. Thus we propose a loss function that while maximizing the inter-class distance and intra-class compactness, allows for the samples which naturally reside further from class center to have a smaller margin. We implement an extremely light and fast to train model using MobileNets and achieve accuracy comparable to state of the art method.
{"title":"Gaussian Soft Margin Angular Loss for Face Recognition","authors":"Bahman Rouhani, Alireza Samadzadeh, M. Rahmati, A. Nickabadi","doi":"10.1109/MVIP49855.2020.9116917","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116917","url":null,"abstract":"Advances in deep learning has lead to drastic improvements in face recognition. A key part of these deep models is their loss function. Consequently developing an efficient and suitable loss function has been an important topic in face recognition in the recent years. Angular-margin-based losses achieve an acceptable performance and inter-class separability. However they are held back by their enforcement of hard margins on all the samples of the training dataset, regardless of whether these samples actually differ from all the other classes enough to enforce a margin. It can be argued that in a large enough dataset with many different settings and age gaps, some faces will look similar to the faces of other classes. In an intuitive and expressive embedding, we expect some faces to be embedded near similar classes with a small margin. Thus we propose a loss function that while maximizing the inter-class distance and intra-class compactness, allows for the samples which naturally reside further from class center to have a smaller margin. We implement an extremely light and fast to train model using MobileNets and achieve accuracy comparable to state of the art method.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114484321","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116901
Saman Hadi, Reza P. R. Hasanzadeh, M. Kersemans
Nowadays composite materials such as carbon fiber reinforced polymers (CFRP)s have been widely used in industrial applications. But, they are susceptible to impact damage and subsequent fatigue cracking and delamination which in long term lead to some negative consequences such as erosion and also breaking the material. Due to the inability to visually observe such defects and also the high sensitivity of industrial components to invasive inspections, non-destructive testing (NDT) techniques are used to deal with the aforementioned problems. In this regards, an ultrasound-based NDT technique called Local defect resonance (LDR) leads to remarkable results for detecting various types of defects in CFRPs. In LDR technique, high frequency acoustical vibrations are used to get a localized resonant activation of a defective region such that these excitation frequencies lead to a significant increase of the vibration amplitude in the defective area relative to the sound area. The problem which arises is that in order to properly localize the defect, the defect resonance frequency must be known which is practically impossible. In this paper, a new defect imaging methodology is proposed, which can localize the defects without any prior knowledge about their location and resonance frequencies. Experiments are performed on a CFRP sample with flat bottom hole (FBH) defects and the proposed method has been quantitatively validated through the experiments by using the signal-to-noise ratio (SNR) criterion. The results show the superiority of our method over some well-known algorithms.
{"title":"A Defect Image Enhancement Approach for Detection of Defective Area in CFRPs Through Local Defect Resonance","authors":"Saman Hadi, Reza P. R. Hasanzadeh, M. Kersemans","doi":"10.1109/MVIP49855.2020.9116901","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116901","url":null,"abstract":"Nowadays composite materials such as carbon fiber reinforced polymers (CFRP)s have been widely used in industrial applications. But, they are susceptible to impact damage and subsequent fatigue cracking and delamination which in long term lead to some negative consequences such as erosion and also breaking the material. Due to the inability to visually observe such defects and also the high sensitivity of industrial components to invasive inspections, non-destructive testing (NDT) techniques are used to deal with the aforementioned problems. In this regards, an ultrasound-based NDT technique called Local defect resonance (LDR) leads to remarkable results for detecting various types of defects in CFRPs. In LDR technique, high frequency acoustical vibrations are used to get a localized resonant activation of a defective region such that these excitation frequencies lead to a significant increase of the vibration amplitude in the defective area relative to the sound area. The problem which arises is that in order to properly localize the defect, the defect resonance frequency must be known which is practically impossible. In this paper, a new defect imaging methodology is proposed, which can localize the defects without any prior knowledge about their location and resonance frequencies. Experiments are performed on a CFRP sample with flat bottom hole (FBH) defects and the proposed method has been quantitatively validated through the experiments by using the signal-to-noise ratio (SNR) criterion. The results show the superiority of our method over some well-known algorithms.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124801679","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116884
Mehran Maneshi, H. Ghassemian, Ghassem Khademi, M. Imani
Technical limitations on the satellite sensors make a trade-off between the spectral and spatial resolution in remotely sensed images. To deal with this issue, pansharpening has been emerged to prepare a single image with the high spatial and spectral resolution, simultaneously. This paper presents a pansharpening approach based on the retina-inspired model and the multiresolution analysis (MRA) framework. The retina- inspired model is simplified by the difference of Gaussian (DoG) operator, and we apply it to the panchromatic image to extract the spatial details. Furthermore, the injection gains in the MRA framework are calculated through an iterative process where the gains at each iteration are updated based on the fusion result obtained from its previous iteration. To investigate the performance of the proposed model, it is compared with some classical pansharpening approaches with two data sets captured by the GeoEye-1 and Pléiades satellite imagery sensors. The experimental results show the proposed retina-inspired pansharpening method acts well in injecting the spatial information along with reducing the spectral distortion.
{"title":"A Retina-Inspired Multiresolution Analysis Framework for Pansharpening","authors":"Mehran Maneshi, H. Ghassemian, Ghassem Khademi, M. Imani","doi":"10.1109/MVIP49855.2020.9116884","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116884","url":null,"abstract":"Technical limitations on the satellite sensors make a trade-off between the spectral and spatial resolution in remotely sensed images. To deal with this issue, pansharpening has been emerged to prepare a single image with the high spatial and spectral resolution, simultaneously. This paper presents a pansharpening approach based on the retina-inspired model and the multiresolution analysis (MRA) framework. The retina- inspired model is simplified by the difference of Gaussian (DoG) operator, and we apply it to the panchromatic image to extract the spatial details. Furthermore, the injection gains in the MRA framework are calculated through an iterative process where the gains at each iteration are updated based on the fusion result obtained from its previous iteration. To investigate the performance of the proposed model, it is compared with some classical pansharpening approaches with two data sets captured by the GeoEye-1 and Pléiades satellite imagery sensors. The experimental results show the proposed retina-inspired pansharpening method acts well in injecting the spatial information along with reducing the spectral distortion.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129626960","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116897
Atefeh Ranjkesh Rashtehroudi, A. Shahbahrami, Alireza Akoushideh
Automated License Plate Recognition (ALPR) has many applications in intelligent transport system. The ALPR has three main steps, License Plate (LP) localization, segmentation and Optical Character Recognition (OCR). Each step needs different techniques in real condition and each technique has its specific characteristics. The LP localization techniques detect the LP after that segmentation algorithms should segment and isolate each character from each other. Finally, the OCR step is applied to recognize the separated characters. The final accuracy depends on the accuracy of each step. To improve the OCR step performance, we combine both segmentation and OCR steps as a single-stage using deep learning techniques such as the You Only Look Once (YOLO) framework. Our experimental results show that this proposed approach recognizes the Iranian LP characters with accuracy 99.2% compared to previous works.
车牌自动识别技术在智能交通系统中有着广泛的应用。ALPR有三个主要步骤:车牌定位、分割和光学字符识别(OCR)。每个步骤在实际条件下都需要不同的技术，每种技术都有其特定的特点。LP定位技术对LP进行检测，然后分割算法对每个字符进行分割和分离。最后，应用OCR步骤对分离字符进行识别。最终的精度取决于每一步的精度。为了提高OCR步骤的性能，我们使用深度学习技术(如You Only Look Once (YOLO)框架)将分割和OCR步骤结合为一个单阶段。实验结果表明，该方法对伊朗语LP字符的识别准确率达到99.2%。
{"title":"Iranian License Plate Recognition using Deep Learning","authors":"Atefeh Ranjkesh Rashtehroudi, A. Shahbahrami, Alireza Akoushideh","doi":"10.1109/MVIP49855.2020.9116897","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116897","url":null,"abstract":"Automated License Plate Recognition (ALPR) has many applications in intelligent transport system. The ALPR has three main steps, License Plate (LP) localization, segmentation and Optical Character Recognition (OCR). Each step needs different techniques in real condition and each technique has its specific characteristics. The LP localization techniques detect the LP after that segmentation algorithms should segment and isolate each character from each other. Finally, the OCR step is applied to recognize the separated characters. The final accuracy depends on the accuracy of each step. To improve the OCR step performance, we combine both segmentation and OCR steps as a single-stage using deep learning techniques such as the You Only Look Once (YOLO) framework. Our experimental results show that this proposed approach recognizes the Iranian LP characters with accuracy 99.2% compared to previous works.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133000564","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116902
Reza Heydari Goudarzi, Seyedeh Somayyeh Mousavi, M. Charmi
Blood Pressure (BP) is one of the most important vital signs of the human body, which its value provides valuable physiological information about cardiac function for physicians. In recent years, many types of research have been done in the field of BP estimation using Photoplethysmography (PPG) signal. On the other hand, the results of the studies on Heart Rate (HR) and Respiration Rate (RR) calculations have been reported using the imaging Photoplethysmography (iPPG) signal. The iPPG signal is a kind of PPG signal that is recorded in a non-contact method using a camera.This paper is among the first studies to provide a new algorithm for estimating BP values using the only iPPG signal and with a non-contact method. The validity of the proposed algorithm was evaluated in a gathered database with 40 people. The algorithm in estimation of the Diastolic Blood Pressure (DBP) was able to achieve mean error of −0.2 and standard deviation of 6.41 mmHg and in estimation of the Systolic Blood Pressure (SBP) was able to achieve mean error of 0.45 and standard deviation of 12.39 mmHg.
{"title":"Using imaging Photoplethysmography (iPPG) Signal for Blood Pressure Estimation","authors":"Reza Heydari Goudarzi, Seyedeh Somayyeh Mousavi, M. Charmi","doi":"10.1109/MVIP49855.2020.9116902","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116902","url":null,"abstract":"Blood Pressure (BP) is one of the most important vital signs of the human body, which its value provides valuable physiological information about cardiac function for physicians. In recent years, many types of research have been done in the field of BP estimation using Photoplethysmography (PPG) signal. On the other hand, the results of the studies on Heart Rate (HR) and Respiration Rate (RR) calculations have been reported using the imaging Photoplethysmography (iPPG) signal. The iPPG signal is a kind of PPG signal that is recorded in a non-contact method using a camera.This paper is among the first studies to provide a new algorithm for estimating BP values using the only iPPG signal and with a non-contact method. The validity of the proposed algorithm was evaluated in a gathered database with 40 people. The algorithm in estimation of the Diastolic Blood Pressure (DBP) was able to achieve mean error of −0.2 and standard deviation of 6.41 mmHg and in estimation of the Systolic Blood Pressure (SBP) was able to achieve mean error of 0.45 and standard deviation of 12.39 mmHg.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126898810","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116880
F. Derikvand, Hassan Khotanlou
In recent years, many segmentation methods have been proposed for brain tumor segmentation, among them deeplearning approaches have good performance and have provided better results than other methods. In this paper, an algorithm based on deep neural networks for segmentation of gliomas tumor is presented which is a combination of different Convolutional Neural Network (CNN) architectures. The proposed method uses local and global features of the brain tissue and consists of preprocessing and post-processing steps which leads to better segmentation. The accuracy of the results was evaluated using the dice score coefficient and the sensitivity on the images obtained from two modalities, Flair and T1, from the BraTs2017 data set and achieved acceptable results compared to state of the art methods.
{"title":"Brain Tumor Segmentation in MRI Images Using a Hybrid Deep Network Based on Patch and Pixel","authors":"F. Derikvand, Hassan Khotanlou","doi":"10.1109/MVIP49855.2020.9116880","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116880","url":null,"abstract":"In recent years, many segmentation methods have been proposed for brain tumor segmentation, among them deeplearning approaches have good performance and have provided better results than other methods. In this paper, an algorithm based on deep neural networks for segmentation of gliomas tumor is presented which is a combination of different Convolutional Neural Network (CNN) architectures. The proposed method uses local and global features of the brain tissue and consists of preprocessing and post-processing steps which leads to better segmentation. The accuracy of the results was evaluated using the dice score coefficient and the sensitivity on the images obtained from two modalities, Flair and T1, from the BraTs2017 data set and achieved acceptable results compared to state of the art methods.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"2014 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130343841","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 : 2020-02-01DOI: 10.1109/MVIP49855.2020.9116876
M. Kavian, A. Nadian-Ghomsheh
Computer vision-based health monitoring systems have gained vast attention especially for physical rehabilitation in the past few years. This paper presents a method for measuring the flexibility of wrist and fingers using leap motion camera. Leap motion was incorporated to acquire the 3D position of hand joints. From the acquired joints, using spatial-temporal features of hand joints, physical exercises targeted at rehabilitating the fingers and wrist range of motion were recognized. Then, appropriate joints selected from the recognized exercises were applied to measure the target range of motion. Apart from the proposed method, the accuracy of leap motion sensor for wrist and fingers range of motion was verified against standard goniometry. Furthermore, the dataset created for this study is published and made publically available for further research in this field. Results of the study showed that leap motion shows promising results for measuring range of motion for several wrist and fingers rehabilitation exercises.
{"title":"Monitoring Wrist and Fingers Range of Motion using Leap Motion Camera for Physical Rehabilitation","authors":"M. Kavian, A. Nadian-Ghomsheh","doi":"10.1109/MVIP49855.2020.9116876","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116876","url":null,"abstract":"Computer vision-based health monitoring systems have gained vast attention especially for physical rehabilitation in the past few years. This paper presents a method for measuring the flexibility of wrist and fingers using leap motion camera. Leap motion was incorporated to acquire the 3D position of hand joints. From the acquired joints, using spatial-temporal features of hand joints, physical exercises targeted at rehabilitating the fingers and wrist range of motion were recognized. Then, appropriate joints selected from the recognized exercises were applied to measure the target range of motion. Apart from the proposed method, the accuracy of leap motion sensor for wrist and fingers range of motion was verified against standard goniometry. Furthermore, the dataset created for this study is published and made publically available for further research in this field. Results of the study showed that leap motion shows promising results for measuring range of motion for several wrist and fingers rehabilitation exercises.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124225753","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}
This study is focused on identifying Persian license plate of Iranian cars in different rain conditions, with different distances and lighting, with simple and complex backgrounds and different angles of stationary cars. A method that is applicable to automated license plate identification systems, which is a type of intelligent transportation system. Systems that have been localized due to the variety of appearance of car license plates in different countries are currently being researched in many countries. Among the important challenges in identifying a vehicle license plate are inappropriate conditions such as adverse weather conditions such as rainy weather, snow, fog and dust, which make it difficult to identify license plates. The proposed method, which is a simple yet efficient method, employs many image processing techniques and morphology operations, and the results of implementing the proposed algorithm in MATLAB 2019b software on 420 Color image of car under low rainfall conditions, moderate rainfall and severe rainfall and storm show accuracy of 81%, 61.5% and 10.5% accuracy in identifying plaque IDs and their separation, respectively.
{"title":"Recognizing Persian Automobile license plates under adverse rainy conditions","authors":"Hossein Rezaei, Maryam Haghshenas, Mahboobehsadat Yasini","doi":"10.1109/MVIP49855.2020.9116886","DOIUrl":"https://doi.org/10.1109/MVIP49855.2020.9116886","url":null,"abstract":"This study is focused on identifying Persian license plate of Iranian cars in different rain conditions, with different distances and lighting, with simple and complex backgrounds and different angles of stationary cars. A method that is applicable to automated license plate identification systems, which is a type of intelligent transportation system. Systems that have been localized due to the variety of appearance of car license plates in different countries are currently being researched in many countries. Among the important challenges in identifying a vehicle license plate are inappropriate conditions such as adverse weather conditions such as rainy weather, snow, fog and dust, which make it difficult to identify license plates. The proposed method, which is a simple yet efficient method, employs many image processing techniques and morphology operations, and the results of implementing the proposed algorithm in MATLAB 2019b software on 420 Color image of car under low rainfall conditions, moderate rainfall and severe rainfall and storm show accuracy of 81%, 61.5% and 10.5% accuracy in identifying plaque IDs and their separation, respectively.","PeriodicalId":255375,"journal":{"name":"2020 International Conference on Machine Vision and Image Processing (MVIP)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126103559","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：