Pub Date : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941129
S. M. Zabihi, Morteza Delgir, H. Pourreza
In this paper, an automated retinal vessel extraction algorithm is represented. A multi-scale morphological algorithm is used for local contrast enhancement of color retinal image. This method enhances vessels not only in color image, but also in the three color components of that image. After feature extraction using LBP and spatial image processing, MLP as a classifier segments the pixels into vessels and non-vessels. Finally, in post processing step, we used binary morphologies for noise removing and smoothing. The performance of the proposed algorithm is tested on the images of DRIVE database.
{"title":"Retinal vessel segmentation using color image morphology and local binary patterns","authors":"S. M. Zabihi, Morteza Delgir, H. Pourreza","doi":"10.1109/IRANIANMVIP.2010.5941129","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941129","url":null,"abstract":"In this paper, an automated retinal vessel extraction algorithm is represented. A multi-scale morphological algorithm is used for local contrast enhancement of color retinal image. This method enhances vessels not only in color image, but also in the three color components of that image. After feature extraction using LBP and spatial image processing, MLP as a classifier segments the pixels into vessels and non-vessels. Finally, in post processing step, we used binary morphologies for noise removing and smoothing. The performance of the proposed algorithm is tested on the images of DRIVE database.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125458342","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941175
M. Khodadadzadeh, Omid Sarrafzade, H. Ghassemian
A new method for automatic classification of traffic signs is proposed in this paper. The proposed method is based on the fusion of colour and morphological information. The strategy consists of three steps. First, colour information in HSI colour space is used to segment the input image and finding the region of interests (ROIs) with red pixels. Then, morphological profile is building by employing opening and closing operators on each band of colour image. Next, statistical feature extraction is performed based on both morphological profile and original colour image. Finally, the feature vector is classified by support vector machines based on one-vs.-rest method. The proposed method was tested on domestic database including four classes of red signs. Experimental results show the hit-rate of about 97% in considerably low process time.
{"title":"A fast traffic sign detection and classification system based on fusion of colour and morphological information","authors":"M. Khodadadzadeh, Omid Sarrafzade, H. Ghassemian","doi":"10.1109/IRANIANMVIP.2010.5941175","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941175","url":null,"abstract":"A new method for automatic classification of traffic signs is proposed in this paper. The proposed method is based on the fusion of colour and morphological information. The strategy consists of three steps. First, colour information in HSI colour space is used to segment the input image and finding the region of interests (ROIs) with red pixels. Then, morphological profile is building by employing opening and closing operators on each band of colour image. Next, statistical feature extraction is performed based on both morphological profile and original colour image. Finally, the feature vector is classified by support vector machines based on one-vs.-rest method. The proposed method was tested on domestic database including four classes of red signs. Experimental results show the hit-rate of about 97% in considerably low process time.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131207503","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941171
M. Shahram Moin, Hamed Rezazadegan Tavakoli, A. Broumandnia, Ieee Senior Member
A new retinal vascular tissue segmentation algorithm, which utilizes Gabor wavelet and local binary patterns, is introduced. It would be shown that how a simple preprocessing step would increase the accuracy of algorithm. Different features have been proposed for retinal vessel detection. One of the most famous features adapted is Gabor wavelet. Thanks to multi-resolution property of Gabor, combination of scales can be used to extract features. However, similar features in feature vector would increase the inter-correlation and may lead to poor result. Also, Local Binary Pattern (LBP) is applied. LBP is a powerful feature for texture analysis. A wise pre-processing strategy is applied to image with regard to feature extraction technique. Contrary to previous methods where a simple pre-processing scheme applied for all feature extraction methods, here each feature extraction will utilize its own suitable preprocessing. It is showed that this enhances the result of segmentation. The proposed method has a low dimension feature vector having only four features. The pre-processing step enhances the results in comparison to a previous method in term of area under the ROC curve The computational results of simulations show the high performance of the proposed method in term of accuracy and speed.
{"title":"A new retinal vessel segmentation method using preprocessed Gabor and local binary patterns","authors":"M. Shahram Moin, Hamed Rezazadegan Tavakoli, A. Broumandnia, Ieee Senior Member","doi":"10.1109/IRANIANMVIP.2010.5941171","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941171","url":null,"abstract":"A new retinal vascular tissue segmentation algorithm, which utilizes Gabor wavelet and local binary patterns, is introduced. It would be shown that how a simple preprocessing step would increase the accuracy of algorithm. Different features have been proposed for retinal vessel detection. One of the most famous features adapted is Gabor wavelet. Thanks to multi-resolution property of Gabor, combination of scales can be used to extract features. However, similar features in feature vector would increase the inter-correlation and may lead to poor result. Also, Local Binary Pattern (LBP) is applied. LBP is a powerful feature for texture analysis. A wise pre-processing strategy is applied to image with regard to feature extraction technique. Contrary to previous methods where a simple pre-processing scheme applied for all feature extraction methods, here each feature extraction will utilize its own suitable preprocessing. It is showed that this enhances the result of segmentation. The proposed method has a low dimension feature vector having only four features. The pre-processing step enhances the results in comparison to a previous method in term of area under the ROC curve The computational results of simulations show the high performance of the proposed method in term of accuracy and speed.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121341632","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941143
S. Mozaffari, Parnia Bahar
This paper focuses on shape description for machine printed Farsi/Arabic subwords retrieval. Fourier descriptor (FD) has been used frequently for shape retrieval applications. In this paper we proposed a simple and effective FD based technique for subwords retrieval. In this method, the small number of global parameters is used to eliminate dissimilar subwords. To investigate the efficiency of the proposed method, it is compared with six common FD signatures on a database including Farsi subwords of 4 fonts and 3 sizes. Experimental results show that the proposed method outperforms the others.
{"title":"Machine printed Farsi/Arabic sub-words retrieval by shape signatures","authors":"S. Mozaffari, Parnia Bahar","doi":"10.1109/IRANIANMVIP.2010.5941143","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941143","url":null,"abstract":"This paper focuses on shape description for machine printed Farsi/Arabic subwords retrieval. Fourier descriptor (FD) has been used frequently for shape retrieval applications. In this paper we proposed a simple and effective FD based technique for subwords retrieval. In this method, the small number of global parameters is used to eliminate dissimilar subwords. To investigate the efficiency of the proposed method, it is compared with six common FD signatures on a database including Farsi subwords of 4 fonts and 3 sizes. Experimental results show that the proposed method outperforms the others.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123540477","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941146
Majid Fakheri, T. Sedghi, M. Amirani
Paper presents a method for object recognition that uses whole images of abstract regions, rather than single regions for classification. A key part of our approach is that we do not need to know where in each image the objects lie. We only utilize the fact that objects exist in an image, not where they are located. We have designed a procedure that learns multivariate models for object classes based on the attributes of abstract regions from multiple segmentations of colour images. The objective of this algorithm is to produce a distribution for each of the object classes being learned.
{"title":"EM segmentation algorithm for colour image retrieval","authors":"Majid Fakheri, T. Sedghi, M. Amirani","doi":"10.1109/IRANIANMVIP.2010.5941146","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941146","url":null,"abstract":"Paper presents a method for object recognition that uses whole images of abstract regions, rather than single regions for classification. A key part of our approach is that we do not need to know where in each image the objects lie. We only utilize the fact that objects exist in an image, not where they are located. We have designed a procedure that learns multivariate models for object classes based on the attributes of abstract regions from multiple segmentations of colour images. The objective of this algorithm is to produce a distribution for each of the object classes being learned.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133638330","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941159
B. Moradi, A. Mirzaei
The main goal in majority of handwriting digit recognition systems is to extract a vector feature for every digit in order to distinguish the digits and classify them in their real classes. In this paper, we propose three different feature extraction methods with kNN classifier for Handwritten Persian/Arabic Digit Recognition. Experiments on real world datasets indicate 2DLDA can provide a solution with improved quality in terms of classification accuracy and computation time performance in contrast to two other methods, PCA and PCA+LDA.
{"title":"Using 2DLDA feature extraction in Handwritten Persian/Arabic Digit Recognition","authors":"B. Moradi, A. Mirzaei","doi":"10.1109/IRANIANMVIP.2010.5941159","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941159","url":null,"abstract":"The main goal in majority of handwriting digit recognition systems is to extract a vector feature for every digit in order to distinguish the digits and classify them in their real classes. In this paper, we propose three different feature extraction methods with kNN classifier for Handwritten Persian/Arabic Digit Recognition. Experiments on real world datasets indicate 2DLDA can provide a solution with improved quality in terms of classification accuracy and computation time performance in contrast to two other methods, PCA and PCA+LDA.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124881687","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941142
R. Akbari, S. Mozaffari
In this paper, we demonstrate that gender estimation technique can increase the accuracy of a face recognition system. If the gender of the input image can be estimated correctly before its recognition and compared only with images of the same sex, errors between males and females during recognition step can be eliminated. Consequently, the accuracy will be boosted. Principal Component Analysis (PCA) face recognition system based on single image has been used in our experiment. To be compatible with this recognizer, the proposed gender estimation algorithm uses also a non-training procedure. A part of FERET database including 292 male and 264 female images has been used. Experimental results show 7% accuracy enhancement for PCA recognition system in the presence of gender estimation.
{"title":"Performance enhancement of PCA-based face recognition system via gender classification method","authors":"R. Akbari, S. Mozaffari","doi":"10.1109/IRANIANMVIP.2010.5941142","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941142","url":null,"abstract":"In this paper, we demonstrate that gender estimation technique can increase the accuracy of a face recognition system. If the gender of the input image can be estimated correctly before its recognition and compared only with images of the same sex, errors between males and females during recognition step can be eliminated. Consequently, the accuracy will be boosted. Principal Component Analysis (PCA) face recognition system based on single image has been used in our experiment. To be compatible with this recognizer, the proposed gender estimation algorithm uses also a non-training procedure. A part of FERET database including 292 male and 264 female images has been used. Experimental results show 7% accuracy enhancement for PCA recognition system in the presence of gender estimation.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122394342","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941161
F. Afsari, E. Eslami
In this paper, a new attempt is being made using Attanassov's intuitionistic fuzzy set theory for image retrieval. Intuitionistic fuzzy sets consider not only membership degree of belonging but also take into account the uncertainty involved in membership degree known as hesitation measure. Color features (expressed in various color representation systems), were intensively used (independently or jointly) during the last decade. We propose to revisit the use of color image contents as image descriptors through the introduction of fuzziness, which naturally arise from the imprecision or vagueness of the pixel color values and human perception. This has been applied in the HSV color space. Hue and value are two color features that are used to construct intuitionistic fuzzy sets; we construct two-dimensional sets which are more suitable than one-dimensional ones. Another key aspect of our method is using fuzzy quantities as a similarity measure between two intuitionistic fuzzy sets instead of a real number due to the imprecision of the similarities. To show the robustness of the proposed method, many experiments with large databases are performed and the results show the high performance of finding similar images.
{"title":"Color image retrieval using intuitionistic fuzzy sets","authors":"F. Afsari, E. Eslami","doi":"10.1109/IRANIANMVIP.2010.5941161","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941161","url":null,"abstract":"In this paper, a new attempt is being made using Attanassov's intuitionistic fuzzy set theory for image retrieval. Intuitionistic fuzzy sets consider not only membership degree of belonging but also take into account the uncertainty involved in membership degree known as hesitation measure. Color features (expressed in various color representation systems), were intensively used (independently or jointly) during the last decade. We propose to revisit the use of color image contents as image descriptors through the introduction of fuzziness, which naturally arise from the imprecision or vagueness of the pixel color values and human perception. This has been applied in the HSV color space. Hue and value are two color features that are used to construct intuitionistic fuzzy sets; we construct two-dimensional sets which are more suitable than one-dimensional ones. Another key aspect of our method is using fuzzy quantities as a similarity measure between two intuitionistic fuzzy sets instead of a real number due to the imprecision of the similarities. To show the robustness of the proposed method, many experiments with large databases are performed and the results show the high performance of finding similar images.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129405023","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941181
M. Alipoor, Z. Ebrahimi, J. Haddadnia
In this paper a novel logarithmic edge detection algorithm is presented. The algorithm is an extended and modified version of PLIP Sobel edge detection algorithm. Six new kernels are suggested to achieve a higher level of independence from scene illumination and provide obvious distinction between edge and non-edge pixels. We present experimental results for this method, and compare results of the algorithm against several leading edge detection methods such as Sobel, Canny and conventional logarithmic edge detection. To compare results objectively, we computed edginess judging index (EJI) for edge detection algorithms. The proposed technique is effective, as demonstrated by computer simulations, conceptually straight forward, and easy to use.
{"title":"A novel logarithmic edge detection algorithm","authors":"M. Alipoor, Z. Ebrahimi, J. Haddadnia","doi":"10.1109/IRANIANMVIP.2010.5941181","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941181","url":null,"abstract":"In this paper a novel logarithmic edge detection algorithm is presented. The algorithm is an extended and modified version of PLIP Sobel edge detection algorithm. Six new kernels are suggested to achieve a higher level of independence from scene illumination and provide obvious distinction between edge and non-edge pixels. We present experimental results for this method, and compare results of the algorithm against several leading edge detection methods such as Sobel, Canny and conventional logarithmic edge detection. To compare results objectively, we computed edginess judging index (EJI) for edge detection algorithms. The proposed technique is effective, as demonstrated by computer simulations, conceptually straight forward, and easy to use.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128948036","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 : 2010-10-01DOI: 10.1109/IRANIANMVIP.2010.5941170
Marzieh Amini, K. Yaghmaie, H. Sadreazami
In this paper, a new halftone image watermarking is presented based on digital wavelet transform combined with singular value decomposition technique. Halftoning is the process of representing grayscale images using just black and white i.e. binary levels. The original image is an error diffusion halftone image which is decomposed into 2-level wavelet transform. In the second level of wavelet transform, the subband with the midst variance intensity is selected as a place for inserting the watermark. The singular value decomposing is applied to this selected subband and watermark image. The embedding modification is done by combining singular value of selected subband with singular value of watermark image. In extraction process, detector response is computed to obtain the original watermark. Experimental results show good robustness against some common signal processing attacks.
{"title":"Error diffusion halftone image watermarking based on SVD-DWT technique","authors":"Marzieh Amini, K. Yaghmaie, H. Sadreazami","doi":"10.1109/IRANIANMVIP.2010.5941170","DOIUrl":"https://doi.org/10.1109/IRANIANMVIP.2010.5941170","url":null,"abstract":"In this paper, a new halftone image watermarking is presented based on digital wavelet transform combined with singular value decomposition technique. Halftoning is the process of representing grayscale images using just black and white i.e. binary levels. The original image is an error diffusion halftone image which is decomposed into 2-level wavelet transform. In the second level of wavelet transform, the subband with the midst variance intensity is selected as a place for inserting the watermark. The singular value decomposing is applied to this selected subband and watermark image. The embedding modification is done by combining singular value of selected subband with singular value of watermark image. In extraction process, detector response is computed to obtain the original watermark. Experimental results show good robustness against some common signal processing attacks.","PeriodicalId":350778,"journal":{"name":"2010 6th Iranian Conference on Machine Vision and Image Processing","volume":"175 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123331967","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：