Pub Date : 2015-03-11DOI: 10.1109/PRIA.2015.7161633
Saber Yaghoobi, Saeed Hemayat, H. Mojallali
Image enhancement methods are known among the most important image processing techniques. Here, image enhancement is considered as an optimization problem and a new heuristic optimization algorithm namely the Black Hole is used to solve it. Image enhancement is a nonlinear optimization problem with its particular constraints and the enhancement process will be done by intensifying each pixel's content. In this paper, BH is employed to find the image's optimum parameters of the transfer function in order to get the best results. BH is used here for its simplicity, ease of implementation, and also its invincibility against the parameter tuning issues. Performance of the proposed enhancement algorithm is tested against some of the well-known enhancement techniques viz. GA, PSO, HE and CS, and the obtained results indicate the robustness and also the outperformance of the proposed algorithm among its other counterparts. Enhancement in opaque images consisting of immense dominant gray values can be listed as one of the proposed method's superiority to that of the other available in literature, which will turn the input image into an enhanced image, featuring embossed textures.
{"title":"Image gray-level enhancement using Black Hole algorithm","authors":"Saber Yaghoobi, Saeed Hemayat, H. Mojallali","doi":"10.1109/PRIA.2015.7161633","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161633","url":null,"abstract":"Image enhancement methods are known among the most important image processing techniques. Here, image enhancement is considered as an optimization problem and a new heuristic optimization algorithm namely the Black Hole is used to solve it. Image enhancement is a nonlinear optimization problem with its particular constraints and the enhancement process will be done by intensifying each pixel's content. In this paper, BH is employed to find the image's optimum parameters of the transfer function in order to get the best results. BH is used here for its simplicity, ease of implementation, and also its invincibility against the parameter tuning issues. Performance of the proposed enhancement algorithm is tested against some of the well-known enhancement techniques viz. GA, PSO, HE and CS, and the obtained results indicate the robustness and also the outperformance of the proposed algorithm among its other counterparts. Enhancement in opaque images consisting of immense dominant gray values can be listed as one of the proposed method's superiority to that of the other available in literature, which will turn the input image into an enhanced image, featuring embossed textures.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127353647","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161646
Ali Katanforoush, Z. Rezvani
This paper aims to stroke recognition, where the strokes are connected forms of cursive handwritten scripts, and in particular, we concern on recognition of Farsi handwriting strokes. In Farsi and some other writing systems, connected letters have special shapes that are often unrecognizable from their separated shapes. Despite that quite efficient algorithms have been developed for recognition of handwritten digits and disjoint letters, adapting these algorithms to stroke recognition is so arduous that development of a holistic approach is preferable. In this paper, we develop a method for Farsi handwriting recognition based on profile-HMM and study aspects of modeling the spatiotemporal features of handwriting strokes. The modular architecture of profile-HMMs provides a flexible framework for stroke modeling. Stroke shrinking and elongation are naturally modeled by the recurrent states and the silent states of profile-HMMs and make the model insensitive to writing speed and subtle slides. Our experimental results show that the profile-HMM is quite robust with respect to downsampling of the curve points, also is robust with respect to various settings in the training procedure. Our method correctly recognizes the main stroke of 90.8%, 98.5%, and 99.2% of handwriting samples, respectively in the top first, top five, and top ten hits.
{"title":"Recognition of Farsi handwriting strokes using profile HMM","authors":"Ali Katanforoush, Z. Rezvani","doi":"10.1109/PRIA.2015.7161646","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161646","url":null,"abstract":"This paper aims to stroke recognition, where the strokes are connected forms of cursive handwritten scripts, and in particular, we concern on recognition of Farsi handwriting strokes. In Farsi and some other writing systems, connected letters have special shapes that are often unrecognizable from their separated shapes. Despite that quite efficient algorithms have been developed for recognition of handwritten digits and disjoint letters, adapting these algorithms to stroke recognition is so arduous that development of a holistic approach is preferable. In this paper, we develop a method for Farsi handwriting recognition based on profile-HMM and study aspects of modeling the spatiotemporal features of handwriting strokes. The modular architecture of profile-HMMs provides a flexible framework for stroke modeling. Stroke shrinking and elongation are naturally modeled by the recurrent states and the silent states of profile-HMMs and make the model insensitive to writing speed and subtle slides. Our experimental results show that the profile-HMM is quite robust with respect to downsampling of the curve points, also is robust with respect to various settings in the training procedure. Our method correctly recognizes the main stroke of 90.8%, 98.5%, and 99.2% of handwriting samples, respectively in the top first, top five, and top ten hits.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116395581","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161622
Atefeh Shagholi, M. Charmi, H. Rakhshan
Human facial video captured with a webcam can be processed to extract the heart rate. The objective of this paper is to show that the distance of the person from the webcam is an important parameter in the accuracy of the estimated heart rate. In fact, with increasing this distance, facial region is limited and the resolution of the video image is diminished which eventually leads to decrease in the accuracy of the estimated heart rate. We have carried out experiments on a data set of 12 subjects. The results of experiment have been compared with the heart rates recorded by a fingertip pulse oximeter in a statistical analysis framework. Our comparison reveals that the root mean square error of the difference of the recorded and the estimated heart rates has been increased from 8.91 for one-half meter away from the webcam to 16.77 for three meters away from the webcam.
{"title":"The effect of the distance from the webcam in heart rate estimation from face video images","authors":"Atefeh Shagholi, M. Charmi, H. Rakhshan","doi":"10.1109/PRIA.2015.7161622","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161622","url":null,"abstract":"Human facial video captured with a webcam can be processed to extract the heart rate. The objective of this paper is to show that the distance of the person from the webcam is an important parameter in the accuracy of the estimated heart rate. In fact, with increasing this distance, facial region is limited and the resolution of the video image is diminished which eventually leads to decrease in the accuracy of the estimated heart rate. We have carried out experiments on a data set of 12 subjects. The results of experiment have been compared with the heart rates recorded by a fingertip pulse oximeter in a statistical analysis framework. Our comparison reveals that the root mean square error of the difference of the recorded and the estimated heart rates has been increased from 8.91 for one-half meter away from the webcam to 16.77 for three meters away from the webcam.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129832848","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161638
Sina Moayed Baharlou, Saeed Hemayat, A. Saberkari, Saber Yaghoobi
A new Persian license plate recognition algorithm is presented. These operations are highly susceptible to error, especially where the image consists of large amount of either vehicle's linked components or the other existing objects. Although the proposed character recognition procedure is highly optimized for Persian plates, the localization parts can be employed for all types of vehicles. Minimum rectangle bounding box is replaced the common bounding box methods, compensating normal bounding box's inherent flaws. License plate possibility ratio (LPPR) is a robust method proposed here to localize the plate. New method of finding plate's location out of so many rectangles, considering “Sensitive to angle” criterions for characters has also been presented. It should be noted that the process is regardless of the plate's location. Different approach on thresholding namely: “Dynamic Thresholding” is used to overcome the probable drawbacks caused by inappropriate lighting. From OCR point of view, a graph, consisting of two specifications will be formed and a set of rules will be defined to capture the character's label. An automated harassment section is added as the denoising filter, in order to omit the grinning ramifications. Presenting the best percent accuracy (95.33%) among relevant well-known algorithms in localization procedure with 25ms run time of the program, and also the outstanding results with over 97% of percent accuracy in character recognition of Persian plates with 30ms run time of the program on Linux and also average of 90ms on Android, can be listed as strong proofs of algorithm's efficiency.
{"title":"Fast and adaptive license plate recognition algorithm for Persian plates","authors":"Sina Moayed Baharlou, Saeed Hemayat, A. Saberkari, Saber Yaghoobi","doi":"10.1109/PRIA.2015.7161638","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161638","url":null,"abstract":"A new Persian license plate recognition algorithm is presented. These operations are highly susceptible to error, especially where the image consists of large amount of either vehicle's linked components or the other existing objects. Although the proposed character recognition procedure is highly optimized for Persian plates, the localization parts can be employed for all types of vehicles. Minimum rectangle bounding box is replaced the common bounding box methods, compensating normal bounding box's inherent flaws. License plate possibility ratio (LPPR) is a robust method proposed here to localize the plate. New method of finding plate's location out of so many rectangles, considering “Sensitive to angle” criterions for characters has also been presented. It should be noted that the process is regardless of the plate's location. Different approach on thresholding namely: “Dynamic Thresholding” is used to overcome the probable drawbacks caused by inappropriate lighting. From OCR point of view, a graph, consisting of two specifications will be formed and a set of rules will be defined to capture the character's label. An automated harassment section is added as the denoising filter, in order to omit the grinning ramifications. Presenting the best percent accuracy (95.33%) among relevant well-known algorithms in localization procedure with 25ms run time of the program, and also the outstanding results with over 97% of percent accuracy in character recognition of Persian plates with 30ms run time of the program on Linux and also average of 90ms on Android, can be listed as strong proofs of algorithm's efficiency.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125742096","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161618
Ladan Sharafyan Cigaroudy, N. Aghazadeh
In this paper, we illustrate an iterative algorithm for extraction of object with tubular structure specially vessel extraction. For this aim, we segment image to reach binary image in which the pixels of purpose object is found. In our segmentation method, we use Gaussian scale-space technique to compute discrete gradient of image for pre-segmenting. Also, in order to denoise, we use tight frame of shearlet transform. This algorithm has an iterative part based on iterative part of TFA [2], but we use eigenvectors of Hessian matrix of image for improving this part. Theoretical properties of this method are presented. The experimental results show that in our algorithm distinguishing homogeneous vessels is done efficiently.
{"title":"A binary-segmentation algorithm based on shearlet transform and eigenvectors","authors":"Ladan Sharafyan Cigaroudy, N. Aghazadeh","doi":"10.1109/PRIA.2015.7161618","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161618","url":null,"abstract":"In this paper, we illustrate an iterative algorithm for extraction of object with tubular structure specially vessel extraction. For this aim, we segment image to reach binary image in which the pixels of purpose object is found. In our segmentation method, we use Gaussian scale-space technique to compute discrete gradient of image for pre-segmenting. Also, in order to denoise, we use tight frame of shearlet transform. This algorithm has an iterative part based on iterative part of TFA [2], but we use eigenvectors of Hessian matrix of image for improving this part. Theoretical properties of this method are presented. The experimental results show that in our algorithm distinguishing homogeneous vessels is done efficiently.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114578018","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}
The ability to decompose superimposed images to their basic components has a fundamental importance in machine vision applications. Segmentation Algorithms consider an image composed of several regions each with a particular gray level, texture or color and try to extract those regions which are not covering each other. However, in this paper, we propose a method for decomposing an image to its superimposed components. Taking prior assumptions into account requires Bayesian framework which is well adapted to this application. Also, a profound mathematical theory called Variational Method is used here which makes us capable of calculating intractable integrals and marginal posteriors. In this paper, situations where superimposed images are to be recovered are discussed and a thorough framework is suggested which is basically founded on the ground of Blind Source Separation (BSS) and Independent Component Analysis (ICA). The main idea of this paper is exerted on some synthetic images to verify its applicability.
{"title":"Separation of multiplicative image components by Bayesian Independent Component Analysis","authors":"Arash Mehrjou, Babak Nadjar Araabi, Reshad Hosseini","doi":"10.1109/PRIA.2015.7161648","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161648","url":null,"abstract":"The ability to decompose superimposed images to their basic components has a fundamental importance in machine vision applications. Segmentation Algorithms consider an image composed of several regions each with a particular gray level, texture or color and try to extract those regions which are not covering each other. However, in this paper, we propose a method for decomposing an image to its superimposed components. Taking prior assumptions into account requires Bayesian framework which is well adapted to this application. Also, a profound mathematical theory called Variational Method is used here which makes us capable of calculating intractable integrals and marginal posteriors. In this paper, situations where superimposed images are to be recovered are discussed and a thorough framework is suggested which is basically founded on the ground of Blind Source Separation (BSS) and Independent Component Analysis (ICA). The main idea of this paper is exerted on some synthetic images to verify its applicability.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130628658","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161642
Saeed Zaravi, R. Amirfattahi, B. Vahdat
2D block method (2D BM) is a new approach to solve inverse problem as one of the most challenging case in Electrical Impedance Tomography (EIT). In this method, a tissue is modeled by some blocks to construct a medical image of a body limb. Recently, a non-iterative linear inverse solution is introduced to solve the inverse problem in the 2D BM. But effect of measurement error has not been considered for non-iterative linear inverse solution yet. In this paper, an appropriate method is proposed to investigate the error propagation in 2D BM. The effect of measurement error is considered as well through different examples. Results show that the BM is very sensitive to the measurement error and fault propagation depends mainly on the type of tissue conductivity distribution. It also can be show that the error increases exponentially in each calculation step.
{"title":"Investigation of error propagation and measurement error for 2D block method in Electrical Impedance Tomography","authors":"Saeed Zaravi, R. Amirfattahi, B. Vahdat","doi":"10.1109/PRIA.2015.7161642","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161642","url":null,"abstract":"2D block method (2D BM) is a new approach to solve inverse problem as one of the most challenging case in Electrical Impedance Tomography (EIT). In this method, a tissue is modeled by some blocks to construct a medical image of a body limb. Recently, a non-iterative linear inverse solution is introduced to solve the inverse problem in the 2D BM. But effect of measurement error has not been considered for non-iterative linear inverse solution yet. In this paper, an appropriate method is proposed to investigate the error propagation in 2D BM. The effect of measurement error is considered as well through different examples. Results show that the BM is very sensitive to the measurement error and fault propagation depends mainly on the type of tissue conductivity distribution. It also can be show that the error increases exponentially in each calculation step.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121544739","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161643
Negar Riazifar, Ehsan Saghapour
Retinal vessel segmentation used for the early diagnosis of retinal diseases such as hypertension, diabetes and glaucoma. There exist several methods for segmenting blood vessels from retinal images. The aim of this paper is to analyze the retinal vessel segmentation based on the clustering algorithm DBSCAN relying on a density-based notion of clusters which is designed to discover clusters of arbitrary shape. DBSCAN requires only one input parameter and a value for this parameter is suggested to the user. The performance of algorithm is compared and analyzed using a number of measures which include sensitivity and specificity. The specificity and sensitivity of this method is 5.36 and 3.82 respectively.
{"title":"Retinal vessel segmentation using system fuzzy and DBSCAN algorithm","authors":"Negar Riazifar, Ehsan Saghapour","doi":"10.1109/PRIA.2015.7161643","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161643","url":null,"abstract":"Retinal vessel segmentation used for the early diagnosis of retinal diseases such as hypertension, diabetes and glaucoma. There exist several methods for segmenting blood vessels from retinal images. The aim of this paper is to analyze the retinal vessel segmentation based on the clustering algorithm DBSCAN relying on a density-based notion of clusters which is designed to discover clusters of arbitrary shape. DBSCAN requires only one input parameter and a value for this parameter is suggested to the user. The performance of algorithm is compared and analyzed using a number of measures which include sensitivity and specificity. The specificity and sensitivity of this method is 5.36 and 3.82 respectively.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"2 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114121928","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161627
A. Rahimi, A. Harati
Image local feature extraction and description is one of the basic problems in computer vision and robotics. However it has still many challenges. On the other hand, in recent years, after the appearance of novel sensors like Kinect camera, RGB-D images are easily available. So it is necessary to extend feature extraction and description methods to be applicable on RGB-D images. In this paper we propose a new approach to feature extraction and description for RGB-D images: Color Gradient Description of Geometrically Stable Regions. The proposed method, first finds smooth regions with uniform changes in surface normal vectors. The process in this stage is inspired from MSER algorithm. Each region then is normalized to a fixed size circle and is rotated toward its dominant orientation to make description affine, scale, and rotation invariant. Finally, color gradients log-polar histogram of normalized regions is used for description. Experimental results show that CGSR features have good performance in illumination and viewpoint changes and outperform state of the art techniques such as SURF and BRAND in matching precision and robustness.
{"title":"CGSR features: Toward RGB-D image matching using color gradient description of geometrically stable regions","authors":"A. Rahimi, A. Harati","doi":"10.1109/PRIA.2015.7161627","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161627","url":null,"abstract":"Image local feature extraction and description is one of the basic problems in computer vision and robotics. However it has still many challenges. On the other hand, in recent years, after the appearance of novel sensors like Kinect camera, RGB-D images are easily available. So it is necessary to extend feature extraction and description methods to be applicable on RGB-D images. In this paper we propose a new approach to feature extraction and description for RGB-D images: Color Gradient Description of Geometrically Stable Regions. The proposed method, first finds smooth regions with uniform changes in surface normal vectors. The process in this stage is inspired from MSER algorithm. Each region then is normalized to a fixed size circle and is rotated toward its dominant orientation to make description affine, scale, and rotation invariant. Finally, color gradients log-polar histogram of normalized regions is used for description. Experimental results show that CGSR features have good performance in illumination and viewpoint changes and outperform state of the art techniques such as SURF and BRAND in matching precision and robustness.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115137993","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 : 2015-03-11DOI: 10.1109/PRIA.2015.7161619
S. M. Tabatabaei, Abdollah Chalechale, Shekoofeh Moghimi
The most expressive manner which human can reveal his emotional states is facial expression. Automatic facial expression recognition is an emerging field of study having extensive applications among which the human-computer interaction (HCI) has received lots of attentions in recent years. The features extracted from facial images, in order to recognize facial expressions, play an essential role in effectiveness of the facial image descriptors. Local binary pattern (LBP) texture descriptors have been known as simple, yet efficient descriptors which are noticeably used for extracting facial patterns from images. Recently, a generalized form of local binary pattern has been introduced which can offer a more precise image description than simple LBP descriptors. Consequently, it would be expected that taking the advantage of using these new LBP texture descriptors will produce more promising results in comparison with use of simple local binary pattern descriptors. In this paper, a novel method has been proposed for image feature extraction using these new image texture descriptors (generalized LBP); then, the obtained results have been compared to the results produced when applying simple LBP descriptors. Furthermore, K-NN and SVM have been used as classifiers in the proposed approach. Finally, a comparison between the proposed method and the existing local binary pattern algorithms for facial expression recognition concludes the superiority of the proposed algorithm over its existing counterparts.
{"title":"Facial expression recognition using high order directional derivative local binary patterns","authors":"S. M. Tabatabaei, Abdollah Chalechale, Shekoofeh Moghimi","doi":"10.1109/PRIA.2015.7161619","DOIUrl":"https://doi.org/10.1109/PRIA.2015.7161619","url":null,"abstract":"The most expressive manner which human can reveal his emotional states is facial expression. Automatic facial expression recognition is an emerging field of study having extensive applications among which the human-computer interaction (HCI) has received lots of attentions in recent years. The features extracted from facial images, in order to recognize facial expressions, play an essential role in effectiveness of the facial image descriptors. Local binary pattern (LBP) texture descriptors have been known as simple, yet efficient descriptors which are noticeably used for extracting facial patterns from images. Recently, a generalized form of local binary pattern has been introduced which can offer a more precise image description than simple LBP descriptors. Consequently, it would be expected that taking the advantage of using these new LBP texture descriptors will produce more promising results in comparison with use of simple local binary pattern descriptors. In this paper, a novel method has been proposed for image feature extraction using these new image texture descriptors (generalized LBP); then, the obtained results have been compared to the results produced when applying simple LBP descriptors. Furthermore, K-NN and SVM have been used as classifiers in the proposed approach. Finally, a comparison between the proposed method and the existing local binary pattern algorithms for facial expression recognition concludes the superiority of the proposed algorithm over its existing counterparts.","PeriodicalId":163817,"journal":{"name":"2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115286126","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：