• Published Issues

    OpenAccess
    • List of Articles E. Kabir

      • Open Access Article

        1 - A New Corner Extraction Method and its Application to Vehicle De
        E. Kabir M. Fathi H. Sadoghi Yazdi
        Corner detection is employed in many areas of image processing and machine vision. Finding all corners, computing the exact position of the corner and robustness of the algorithm against noise are important criteria in corner detection. In this paper, using the singular More
        Corner detection is employed in many areas of image processing and machine vision. Finding all corners, computing the exact position of the corner and robustness of the algorithm against noise are important criteria in corner detection. In this paper, using the singular values of the matrix defined on the gradient of a small area of the image, a suitable corner is extracted. The proposed method in comparison with the computational method which is based on the eigenvalues of the cross correlation matrix of the gradient of image shows a better performance. It also yields good results in the presence of noise. These two methods were compared on the synthesized and real images of a traffic scene. The proposed method presented better results. Manuscript profile
      • Open Access Article

        2 - A Two Step Method for the Recognition of Printed Subwords
        E. Kabir A. ebrahimi
        In this paper a two step method for the recognition of printed subwords is proposed. Using characteristic loci features, the set of printed subwords are clustered into 300 clusters by k-means algorithm. Each cluster is represented by its mean. In the first step, each in More
        In this paper a two step method for the recognition of printed subwords is proposed. Using characteristic loci features, the set of printed subwords are clustered into 300 clusters by k-means algorithm. Each cluster is represented by its mean. In the first step, each input is classified into 300 categories by minimum Euclidian distance from the cluster centers, and 10 closest clusters are found. In the second step, Fourier descriptors of the subword contour are used to classify the input subword into the members of these 10 clusters. The training set consists of 12700 Farsi subwords in 4 different fonts, Lotus, Mitra, Yagut and Zar, and 3 sizes of 10, 12 and 14. In a test, a set of 500 subwords was used. Considering the first class, top five and top ten classes, 71.4%, 95%, and 98.2% of these subwords were correctly classified. In the post processing, dots of the subword and their positions were used to improve the recognition results. This improved the recognition rate to 92.6%. Manuscript profile
      • Open Access Article

        3 - Top-Down Tracking Algorithm Based on Vehicle Trajectory Learning in the Traffic Scene
        H. Sadoghi Yazdi M. Lotfizad M. Fathy E. Kabir
        In this paper, a trajectory learning-based vehicle tracking algorithm is presented which is a new top-down vehicle tracker. The history of trajectory is learnt by a novel sptio-temporal data base known center transition matrix, CTM. At first, the CTM is constructed on c More
        In this paper, a trajectory learning-based vehicle tracking algorithm is presented which is a new top-down vehicle tracker. The history of trajectory is learnt by a novel sptio-temporal data base known center transition matrix, CTM. At first, the CTM is constructed on centers which are obtained using fuzzy clustering on vehicle trajectories. The i, j-th element of CTM indicates passing of the object from center i to center j in two consecutive frames which CTM is completed by multi-object tracking. The CTM is efficient in search of similar blobs in image sequences and can determine the radius and region of search and increasing of convergence rate of RLS predictor. The proposed tracking algorithm is tested in the intersection of a highway to a square which gives good results. Manuscript profile
      • Open Access Article

        4 - A Simple Method for Online Recognition of Farsi Subwords
        S. M. Razavi E. Kabir
        In this paper, a method for online recognition of Farsi subwords is presented. First, the dots and other signs of the input subword and their relative locations are recognized and the related group to that subword is chosen. If there is only one member in that group, it More
        In this paper, a method for online recognition of Farsi subwords is presented. First, the dots and other signs of the input subword and their relative locations are recognized and the related group to that subword is chosen. If there is only one member in that group, its class is assigned to the input subword, otherwise, the subword body is compared to those of the group members and the subword with minimum distance to the input subword is found. The recognition system also proposes a maximum of nine subwords in the next ranks. The proposed method was tested on a database of 11 samples of 1000 subwords from different writers. The correct recognition rate for the first choice was 74.95%. It reached to 97.87% for the top 10 choices. Manuscript profile
      • Open Access Article

        5 - Computer Aided Graphology for Farsi Handwriting
        A. A. Bahrami Sharif E. Kabir
        Graphology is the science of study and analysis of the personality of an individual from his/her style of handwriting. In western communities, the most important application of graphology is the recruitment of job applicants. In this regard, computer aided extraction an More
        Graphology is the science of study and analysis of the personality of an individual from his/her style of handwriting. In western communities, the most important application of graphology is the recruitment of job applicants. In this regard, computer aided extraction and analysis of features from handwriting can be of great assistance to graphologists. The most dominant features of handwriting employed in graphology include the shape of the page margins, line spacing, line skew, word slant, size of letters, text density, writing speed and regularity. In this paper a number of methods are proposed for automated extraction of some of these features from Farsi handwriting. Experimental results on 118 test samples of different writers are presented and discussed. Manuscript profile
      • Open Access Article

        6 - A New Ensemble Learning Method for Improvement of Classification Performance
        S. H. Nabavi-Kerizi E. Kabir
        The combination of multiple classifiers is shown to be suitable for improving the performance of pattern recognition systems. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. The methods have been proposed for dive More
        The combination of multiple classifiers is shown to be suitable for improving the performance of pattern recognition systems. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. The methods have been proposed for diversity creation can be classified into implicit and explicit methods. In this paper, we propose a new explicit method for diversity creation. Our method adds a new penalty term in learning algorithm of neural network ensembles. This term for each network is the product of its error and the sum of other networks errors. Experimental results on different data sets show that proposed method outperforms the independent training and the negative correlation learning methods. Manuscript profile
      • Open Access Article

        7 - Defect Detection in Textile Fabrics Using Modified Local Binary Patterns
        F. Tajeripour E. Kabir a. sheikhi
        One of the methods which can produce powerful features for texture classification is Local Binary Patterns, LBP. In this paper we propose a method for defect detection in textile fabrics using these features. In the training stage, at first step LBP operator is applied More
        One of the methods which can produce powerful features for texture classification is Local Binary Patterns, LBP. In this paper we propose a method for defect detection in textile fabrics using these features. In the training stage, at first step LBP operator is applied to an image of defect free fabric, pixel by pixel, and the reference feature vector is computed. Then this image is divided into windows and LBP operator is applied on each of these windows. Based on comparison to the reference feature vector a suitable threshold for defect free windows is found. In the detection stage, a test image is divided into windows and using the threshold, defective windows can be detected. The proposed method is gray scale and shift invariant and can be used for defect detection in patterned and plain fabrics. Due to its simplicity online implementation is possible. Manuscript profile
      • Open Access Article

        8 - Detection of Surface Defects on Apples for Quality Grading
        M. Bazhan E. Kabir
        In this paper, two kinds of defects in Golden Delicious apples are recognized: bruise and russet. Russet is divided to two classes: russet in stem-end and russet out of stem-end. Apples are graded into three classes I, II and rejected, according to European standard. To More
        In this paper, two kinds of defects in Golden Delicious apples are recognized: bruise and russet. Russet is divided to two classes: russet in stem-end and russet out of stem-end. Apples are graded into three classes I, II and rejected, according to European standard. To grade the apples, it is necessary to classify apple images into six classes: stem, calyx, bruise, russet in stem-end, russet out of stem-end and healthy. In this method, after pixel-based classification based on RGB color features by a perceptron neural network, correction in classification and stem detection is made. Hue and saturation features are used to correct the image regions classified to bruise. The correction of regions classified to calyx, russet in stem-end and russet out of stem-end is made based on the distance from the gravity center of the stem to the gravity center of each region. This paper presents a new method for defect classification and sub classification of russet to two classes, russet in stem-end and russet out of stem-end. Experimental results of the proposed algorithm show that the correct grading rate of 120 apple images is 81.66%. The grading errors result from misdetection of stem and errors in defect detection. Manuscript profile
      • Open Access Article

        9 - A Two-Stage Method for Classifiers Combination
        S. H. Nabavi Karizi E. Kabir
        Ensemble learning is an effective machine learning method that improves the classification performance. In this method, the outputs of multiple classifiers are combined so that the better results can be attained. As different classifiers may offer complementary informat More
        Ensemble learning is an effective machine learning method that improves the classification performance. In this method, the outputs of multiple classifiers are combined so that the better results can be attained. As different classifiers may offer complementary information about the classification, combining classifiers, in an efficient way, can achieve better results than any single classifier. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. In this paper, we propose a two-stage method for classifiers combination. In the first stage, by mixture of experts strategy we produce different classifiers and in the second stage by using particle swarm optimization (PSO), we find the optimal weights for linear combination of them. Experimental results on different data sets show that proposed method outperforms the independent training and mixture of experts methods. Manuscript profile
      • Open Access Article

        10 - Design and Implementation of Two Pipeline Architectures for Computing High-Order Moments of Grey-Level Images
        M. Monajati E. Kabir  
        Moments are utilized in image processing for pattern recognition, machine vision and numerous feature extraction techniques. Due to computational complexity, it is difficult to use high order moments in real time processing. This paper presents the design of two new arc More
        Moments are utilized in image processing for pattern recognition, machine vision and numerous feature extraction techniques. Due to computational complexity, it is difficult to use high order moments in real time processing. This paper presents the design of two new architectures for real time computation of moments, up to order 14, M00 to M77, in gray level images, based on parallel systolic arrays and pipelining technique, using a 0.18μm CMOS technology. Implementation of the moment processing element (MPE) of the first architecture illustrates a processing speed of 125 frames/s for 1024×1024 grey-level images. The maximum operating frequency and the power consumption for an architecture with 5 elements is 133 MHz and 14.36 mW, respectively. Since the design is very low power, the number of parallel MPE’s can be easily increased. Simulation shows that with 11 parallel MPE’s, the first 49 moments of 1024×1024 image are computed with the speed of 30 frames/sec. To further decrease the latency of the first architecture, the second architecture is proposed, in which the add operation is performed only with a single adder and a compressor. Simulation shows that the latency of the second architecture is 3.3 times lower than that of the first architecture. Implementation of the second architecture illustrates the maximum operating frequency and the power consumption of 125 MHz and 58.34 mW, respectively. Operating frequency and power consumption of the second architecture is approximately the same as that of the first architecture which befit real time applications. Manuscript profile
      • Open Access Article

        11 - Design of Low Power High Speed Dilation Operator for Binary Images in CMOS Technology
        M. hajirahimi E. Kabir  
        This paper describes the design of hybrid wave-pipeline architecture for implementation of real time morphological dilation. With minor changes to this architecture, it can be utilized for erosion, closing, and opening operators. The new architecture results in higher s More
        This paper describes the design of hybrid wave-pipeline architecture for implementation of real time morphological dilation. With minor changes to this architecture, it can be utilized for erosion, closing, and opening operators. The new architecture results in higher speed, less hardware complexity, and lower area and power dissipation compared to conventional pipeline implementation. In addition, it is faster than the wave-pipeline structure, without the difficulty of balancing the delay of long signal paths. Using the new architecture, three ASIC chips in 0.18µm CMOS are designed for binary image processing through Verilog. These chips dilate a 1024×1024 image by a 21×21 structuring element in 256.58μ s. The maximum frequency of the operations is 5.882 GHz, 5 GHz, and 4.167 GHz. For the power supply of 1.8 V and the 4.167 GHz frequency, the power dissipation is 597mW, 478 mW, and 410 mW, and the chip area is 0.118 mm2, 0.087 mm2, and 0.075 mm2, respectively. Manuscript profile
      • Open Access Article

        12 - Evaluating Two Approaches for Farsi OCR Based on Sub-Word Shape Recognition
        H. Khosravi E. Kabir
        Two approaches for the recognition of printed Farsi documents based on sub-word shape recognition is proposed. First approach is based on recognition of sub-word shape as a whole and the second is based on the recognition of the body of sub-words. Sub-word body is const More
        Two approaches for the recognition of printed Farsi documents based on sub-word shape recognition is proposed. First approach is based on recognition of sub-word shape as a whole and the second is based on the recognition of the body of sub-words. Sub-word body is constructed via removing dots and signs of the sub word. In second approach, information of dots and signs will be added after recognition of the body. Both approaches have two phases: training and test. In training phase, sub-words are clustered based on ISODATA algorithm. Initial centers of the clusters are computed through a hierarchical clustering algorithm. In first approach, sub-word recognition is performed in two stages: finding clusters close to the input sub-word and then finding the best match within the sub-words of these clusters. In the second approach another stage is required to find the final sub-word including dots and signs. Experimental results show that on clean images the first algorithm have better performance; 94% versus 93% in word level. But when dealing with low quality and noisy images, both algorithms are suffering from reduced accuracy. Sometimes this reduction is significant. The reasons of this behavior are inspected and some solutions are presented. Finally we compared both methods and inspected pros and cons of Farsi OCR based on sub-word shape. Manuscript profile
      • Open Access Article

        13 - Online Signature Verification in Stationary Wavelet Transform Domain
        M. Valizadeh E. Kabir
        In this paper, an online signature verification method using extended regression in stationary wavelet domain is presented. To calculate the similarity between two signatures by extended regression, we should equalize the time length of the corresponding signals in two More
        In this paper, an online signature verification method using extended regression in stationary wavelet domain is presented. To calculate the similarity between two signatures by extended regression, we should equalize the time length of the corresponding signals in two signatures. Using all points of the signals to equalize their time length will decrease the difference between a genuine signature and its forgery. Here a new approach based on the extreme points warping of the signals is presented. This approach equalizes the time length of two signals without degrading the differences between them. Also we calculated the similarity of signatures by using the details of the signals in stationary wavelet transform, SWT, domain, which showed very good results. The proposed system was tested on SVC2004 signature database. The results were compared with the results of participant teams in the first international signature verification competition. We have gained EER=6% for skilled forgery signatures. Comparing the result, it shows that we stand in the second rank between all the participants. This system has no verification error for random forgery signatures and stands in the first rank. Our experimental results show that using SWT domain instead of time domain decreases the verification error rate by 35%. Manuscript profile
      • Open Access Article

        14 - A Contrast Independent Algorithm for Binarization of Document Images
        M. Valizadeh E. Kabir
        In this paper, we present a contrast independent algorithm for binarization of degraded document images. The proposed algorithm does not require any parameter setting by user. Therefore, it can handle document images with variable foreground and background intensities a More
        In this paper, we present a contrast independent algorithm for binarization of degraded document images. The proposed algorithm does not require any parameter setting by user. Therefore, it can handle document images with variable foreground and background intensities and low contrast documents. The proposed algorithm involves three consecutive stages. At the first stage, independent of contrast between foreground and background, sensible parts of each character are extracted using the modified water flow model, which is designed for the extraction of sensible part of each character and the drawbacks of water flow model are solved in this algorithm. In the second stage, the gray levels of foreground are estimated using the extracted text pixels and the gray levels of background are locally estimated by averaging the original image. At the third stage, for each pixel of image, the average of estimated foreground and background gray levels is defined as local threshold. After extensive experiments, the proposed binarization algorithm demonstrates superior performance against conventional binarization algorithms on a set of degraded document images captured with camera. Proposed algorithm efficiently extracts the low contrast texts. Manuscript profile
      • Open Access Article

        15 - A Method for Automatic Printing Carpet Map Reading and Comparing to C-Means Clustering
        Ahmad Izadipour E. Kabir
        The subject of this paper is to read carpet pattern automatically by computer. This is composed of two steps: detection of vertical and horizontal lines in the pattern and color reduction. Color reduction is essential because of limitation of the number of colors that i More
        The subject of this paper is to read carpet pattern automatically by computer. This is composed of two steps: detection of vertical and horizontal lines in the pattern and color reduction. Color reduction is essential because of limitation of the number of colors that is used in a carpet. To accomplish of this process, we must detect the grid lines on the carpet pattern automatically. These lines are two types: thin lines and thick lines. At the first stage, the distance between thin lines is obtained. Having the first thin line detected, the other thin lines are drawn using this distance. We use a Comb method for detection of thick lines. The major problem in line detection is lagging or leading of the lines due to the mismatch between sampling frequency of the scanner and image resolution. We compensate this distortion in various steps in our algorithm. In the second step, we want all the pixels in the same square, to have the same color. This is obtained by mapping colors to the best color in the palette. We propose three methods. In first method the user selects two selections per any colors. Palette is obtained from some processes in these selections. Those pixels that are in the middle of the squares are mapped to the palette. Then color histogram is computed. The color that has the maximum histogram value is assigned to the square. In order to decrease user’s interference, C-means clustering algorithm is used in two types. The centers of initial clusters are determined once with user’s interference and once randomly. Results of these three methods are compared. We tested our methods on 20 samples of carpet patterns, and the error rate was variable from 0.07% to 0.5% between samples. Manuscript profile
      • Open Access Article

        16 - Analysis of Supervised Learners to Extract Knowledge of Lighting Angels in Face Images
        S. Naderi N. Moghadam Charkari E. Kabir
        Variation of Light intensity and its direction have been the main challenges in many face recognition systems that lead to the different normal and abnormal shadows. Today, various methods are presented for face recognition under different lighting conditions which requ More
        Variation of Light intensity and its direction have been the main challenges in many face recognition systems that lead to the different normal and abnormal shadows. Today, various methods are presented for face recognition under different lighting conditions which require previous knowledge about Light source and the angle of radiation as well. In this paper, a new approach is proposed to extract the knowledge of/about the lighting angle/direction in face images based on learning techniques. At First, some effective coefficients on lighting variation are extracted on DCT domain. They will be used to determine lighting classes after normalization. Then, three different learning algorithms, Decision tree, SVM, and WAODE (Weightily Averaged One-Dependence Estimators) are used to learn the lighting classes. The algorithms have been tested on the well known YaleB and Extended Yale face databases. The comparative results indicate that the SVM achieves the best average accuracy for classification. On the other hand, WAODE Bayesian approach attains the better accuracy in classes with large lighting angle because of its resistance against data loss. Manuscript profile
      • Open Access Article

        17 - Color reduction for Machine-Printed Carpet Pattern by Reinforcement Learning
        M. Fateh E. Kabir M. Nili Ahmadabadi
        Automatic reading of carpet patterns Requires To find the original colors of the pattern in a scanned image. It includes detecting of pattern lines and reducing the number of colors in the image. Color reduction is done in two steps: Finding the best pallet and mapping More
        Automatic reading of carpet patterns Requires To find the original colors of the pattern in a scanned image. It includes detecting of pattern lines and reducing the number of colors in the image. Color reduction is done in two steps: Finding the best pallet and mapping the image colors to the pallet colors. The accuracy of color reduction is so important that it may be required to ask for user intervention. The purpose of this study is to provide a new method in automatic color reduction with high accuracy. To achieve this target, reinforcement learning method is used which yields a 98% accuracy. This is a new method in color reduction and no one has used it yet. This method is defined with respect to the application and the amount of color reduction is such that does not degrade the accuracy. Therefore, the resulting pallet has more colors comparing to the original one. In the work reported in this article, first the grid lines of the pattern are detected. Then a single color is assigned to each box of the grid. After these steps, through the reinforcement learning method the color reduction is carried out. The results obtained from applying the proposed algorithm on some sample images are reported and discussed. Manuscript profile
      • Open Access Article

        18 - Sub-Word Image Clustering in Old Printed Documents Using Template Matching
        M. R. Soheili E. Kabir
        Due to the rapid growth of digital libraries, digitizing large documents has become an important topic. In a quite long book, similar characters, sub-words and words will occur many times. In this paper, we propose a sub-word image clustering method for the applications More
        Due to the rapid growth of digital libraries, digitizing large documents has become an important topic. In a quite long book, similar characters, sub-words and words will occur many times. In this paper, we propose a sub-word image clustering method for the applications dealing with large uniform documents. We assumed that the whole document is printed in a single font and print quality is not good. To test our method, we created a dataset of all sub-words of a Farsi book. The book has 233 pages with more than 111000 sub-words manually labeled. We use an incremental clustering algorithm. Four simple features are extracted from each sub-word and compared with the corresponding features of each cluster center. If all features' differences lie within certain thresholds, the sub-word and the winner cluster center are finely compared using a template matching algorithm. In our experiments, we show that all sub-words of the book are recognized with more than 99.7% accuracy by assigning the label of each cluster center to all of its members. Manuscript profile
      • Open Access Article

        19 - Using Prominent Regions in Search Space Reduction for Recognition of Printed Farsi Subwords
        H. Davoudi E. Kabir
        In the most common Lexicon reduction methods, lexicon words are clustered based on their holistic shape features and then each query word image is classified into the closest cluster. As the errors at this stage propagate to the subsequent stages, relevant clusters shou More
        In the most common Lexicon reduction methods, lexicon words are clustered based on their holistic shape features and then each query word image is classified into the closest cluster. As the errors at this stage propagate to the subsequent stages, relevant clusters should be selected with a high degree of accuracy. In this paper we present a novel verification method which decides on the validity of the recognized clusters based on a proposed confidence measure. The level of confidence to the selected clusters is measured using local shape features in the verification phase, where it is determined that the selected cluster is acceptable or not. For this purpose, some local shape features of the input subword image are compared to the “prominent regions” of the corresponding cluster. The prominent regions of a cluster are some local regions that discriminate the members of that cluster compared to the other clusters. The proposed verification method along with some predefined rules is used to reduce the lexicon size of Farsi subwords. The experiments conducted on a set of 6895 common Farsi subwords show that our proposed method significantly reduces the search space while preserving the accuracy in an acceptable rate. Manuscript profile
      • Open Access Article

        20 - Color Reduction of Hand-painted Carpet Patterns Before Discretization
        M. Fateh E. Kabir
        Carpet patterns are in two categories: machine-painted and hand-painted. Hand-painted patterns are divided into two groups: before and after discretization. The purpose of this study is color reduction of hand-painted patterns before discretization. There are some artic More
        Carpet patterns are in two categories: machine-painted and hand-painted. Hand-painted patterns are divided into two groups: before and after discretization. The purpose of this study is color reduction of hand-painted patterns before discretization. There are some articles about color reduction of hand-painted carpet patterns after discretization, but so far, there is not an article on patterns before discretization. The proposed algorithm consists of the following steps: image segmentation, finding the color of each region, color reduction around the edges and final color reduction with C-means. For 80 segments of different 20 patterns, the algorithm has an approximate of 96% accuracy. In other words, the colors of 96% of image pixels are found correctly. The high accuracy of this method is due to its fitness to the application. The proposed method is not fully automatic and requires the total number of colors as its input. Manuscript profile
      • Open Access Article

        21 - Fusion of Neural Networks Based on Negative Correlation Learning for Offline Handwritten Word Recognition
        S. A. A. Abbaszadeh Arani E. Kabir
        In this study, an ensemble classification method, based on negative correlation learning, is used for holistic recognition of handwritten words with limited vocabulary. In this method, training data set, after preprocessing and feature extraction, is applied to the base More
        In this study, an ensemble classification method, based on negative correlation learning, is used for holistic recognition of handwritten words with limited vocabulary. In this method, training data set, after preprocessing and feature extraction, is applied to the base Multilayer Perceptron classifiers. These classifiers are trained by negative correlation learning to make them diverse. Features extracted from a test input are applied to the base classifiers, which produce somehow diverse outputs. By combining these outputs, the final output of the system is obtained. For experiments, three feature sets based on zoning, gradient image and contour chain code are extracted from the images. In experiments, performed on 775 images of 31 Province centers from "Iranshahr" dataset, when gradient-based features were used to train 6 Multilayer Perceptron classifiers by negative correlation, by Fusion the outputs of these classifiers through voting, an average recognition rate of 96.10 percent is achieved. Manuscript profile
      • Open Access Article

        22 - Onset Detection for Tar Solo Based on Pitch and Energy Features
        B. Farrokhi E. Kabir
        This paper develops a new method of onset detection for the Tar, a traditional Iranian musical instrument. The proposed method is based on both types of pitch and energy features and an adaptive peak picking algorithm is utilized for primary onset detection. An improved More
        This paper develops a new method of onset detection for the Tar, a traditional Iranian musical instrument. The proposed method is based on both types of pitch and energy features and an adaptive peak picking algorithm is utilized for primary onset detection. An improved template matching method is used to detect fundamental frequencies and finally, onsets are tagged based on primary onsets and fundamental frequencies. This step is especially useful to detect the reaz, repeatedly played notes with the same frequency and short durations. For the evaluation of the method, a data set with predetermined onsets was produced and the results were compared with an energy based method explained in terms of F measure. Manuscript profile
      • Open Access Article

        23 - Example-Based single Document Image Super Resolution Using Asynchronous Sequential Gradient Descent Algorithm
        A. Abedi E. Kabir
        In this paper, a new method for resolution enhancement of single document images is presented. The proposed method is example based using an example set of low-resolution and high-resolution training patches. According to the Bayes rule, one function is considered as th More
        In this paper, a new method for resolution enhancement of single document images is presented. The proposed method is example based using an example set of low-resolution and high-resolution training patches. According to the Bayes rule, one function is considered as the likelihood or data-fidelity term that measures the fidelity of the output high-resolution to the input low-resolution image. As well, three other functions are considered as the regularization terms containing the prior knowledge about the desired high-resolution document image. Three priors which are fulfilled by the regularization terms are bimodality of document images, smoothness of background and text regions, and similarity to the patches in the example set. By minimizing these four energy functions through the iterative procedure of asynchronous sequential gradient descent, the HR image is reconstructed. Instead of synchronous minimization of the linear combination of these functions, they are minimized in order and according to the gradual changes in their values and in the updating HR image. Therefore, determining the coefficients of the linear combination, which are variable for input images, is no longer required. In the experimental results on twenty document images with different fonts, at different resolutions, and with different amounts of noise and blurriness, the proposed method achieves significant improvements in visual image quality and in reducing the computational complexity. Manuscript profile
      • Open Access Article

        24 - Presenting Technique for the Quantitative Evaluation of Image Color Reduction Algorithms by Explaining a Practical Sample
        M. Fateh E. Kabir
        In color reduction algorithms the result will be evaluated based on visual or qualitative standards. Evaluation without considering the quantitative standard wouldn't be a complete and accurate evaluation and trends of viewer are very effective on the evaluation. In som More
        In color reduction algorithms the result will be evaluated based on visual or qualitative standards. Evaluation without considering the quantitative standard wouldn't be a complete and accurate evaluation and trends of viewer are very effective on the evaluation. In some articles, the result will be evaluated with MSE. In this standard error the difference between the final images’ pixels color with first image will be considered as a failure in which is not a suitable technique for evaluating of color reduction methods. In images color reduction, if a color completely be replaced by a color closed to the original color it wouldn’t be considered as a failure. If these replacements don’t happen for all of those specific color pixels, then an error has happened in color reduction. The disintegration of the resulted colors from color reduction algorithm with desired colors should be considered in presenting the evaluation criteria since this will not be considered in MSE. In some of color reduction applications such as color reduction in the carpet cartoons, the final desired pixel color is specified and presenting the wrong color will be an error. Therefore, in such applications, the quantitative evaluation based on final color of each pixel is possible. By presenting criteria for quantitative evaluation, viewer trends wouldn't be considered in evaluation and the possibility of accurate comparison of color reduction algorithms would take place. In this article, we have presented a technique of quantitative evaluation for color reduction algorithms. When the final desired color for pixels are specified, this criteria would work out. To demonstrate the functionality of this quantitative evaluation technique, one of the applications of color reduction which is color reduction in carpet cartoons would be discussed. Several methods of color reduction would be evaluated based on proposed evaluation criteria and reference [42], had the lowest error. Manuscript profile