Showing papers in "International Journal of Computer Science and Information Technology in 2011"
TL;DR: In this paper an attempt is made to study the performance of most commonly used edge detection techniques for image segmentation and the comparison of these techniques is carried out with an experiment by using MATLAB software.
Abstract: Interpretation of image contents is one of the objectives in computer vision specifically in image processing. In this era it has received much awareness of researchers. In image interpretation the partition of the image into object and background is a severe step. Segmentation separates an image into its component regions or objects. Image segmentation t needs to segment the object from the background to read the image properly and identify the content of the image carefully. In this context, edge detection is a fundamental tool for image segmentation. In this paper an attempt is made to study the performance of most commonly used edge detection techniques for image segmentation and also the comparison of these techniques is carried out with an experiment by using MATLAB software.
420 citations
TL;DR: A generic computer forensics investigation model, known as GCFIM is proposed, based on the commonly shared processes, which would make it easier for the new users to understand the processes and to serve as the basic underlying concept for the development of a new set of processes.
Abstract: The increasing criminal activities using digital information as the means or targets warrant for a structured manner in dealing with them. Since 1984 when a formalized process been introduced, a great number of new and improved computer forensic investigation processes have been developed. In this paper, we reviewed a few selected investigation processes that have been produced throughout the years and then identified the commonly shared processes. Hopefully, with the identification of the commonly shard process, it would make it easier for the new users to understand the processes and also to serve as the basic underlying concept for the development of a new set of processes. Based on the commonly shared processes, we proposed a generic computer forensics investigation model, known as GCFIM.
136 citations
TL;DR: An off-line handwritten alphabetical character recognition system using multilayer feed forward neural network is described in this article, where a new method, called, diagonal based feature extraction is introduced for extracting the features of the handwritten alphabets.
Abstract: An off-line handwritten alphabetical character recognition system using multilayer feed forward neural network is described in the paper. A new method, called, diagonal based feature extraction is introduced for extracting the features of the handwritten alphabets. Fifty data sets, each containing 26 alphabets written by various people, are used for training the neural network and 570 different handwritten alphabetical characters are used for testing. The proposed recognition system performs quite well yielding higher levels of recognition accuracy compared to the systems employing the conventional horizontal and vertical methods of feature extraction. This system will be suitable for converting handwritten documents into structural text form and recognizing handwritten names.
102 citations
TL;DR: Some of the most popular machine learning methods and their applicability to the problem of spam Email classification are reviewed and a comparison of their performance on the SpamAssassin spam corpus is presented.
Abstract: The increasing volume of unsolicited bulk e-mail (also known as spam) has generated a need for reliable anti-spam filters. Machine learning techniques now days used to automatically filter the spam e-mail in a very successful rate. In this paper we review some of the most popular machine learning methods (Bayesian classification, k-NN, ANNs, SVMs, Artificial immune system and Rough sets) and of their applicability to the problem of spam Email classification. Descriptions of the algorithms are presented, and the comparison of their performance on the SpamAssassin spam corpus is presented.
86 citations
TL;DR: Teaching strategies matching with learner’s personality using the Myers-Briggs Type Indicator (MBTI) tools and the result reveals the system effectiveness for which it appears that the proposed approach may be promising.
Abstract: Personalized e-learning implementation is recognized one of the most interesting research areas in the distance web-based education. Since the learning style of each learner is different we must to fit elearning to the different needs of learners. This paper discusses teaching strategies matching with learner’s personality using the Myers-Briggs Type Indicator (MBTI) tools. Based on an innovative approach, a framework for building an adaptive learning management system by considering learner’s preference has been developed. The learner’s profile is initialized according to the results obtained by the student in the index of learning styles questionnaire and then fine-tuned during the course of the interaction using the Bayesian model. Moreover, an experiment was conducted to evaluate the performance of our approach. The result reveals the system effectiveness for which it appears that the proposed approach may be promising.
71 citations
TL;DR: This paper proposes a real time mobile health system for monitoring elderly patients from indoor or outdoor environments that uses a bio- signal sensor worn by the patient and a Smartphone as a central node.
Abstract: Recent research in ubiquitous computing uses technologies of Body Area Networks (BANs) to monitor the person's kinematics and physiological parameters. In this paper we propose a real time mobile health system for monitoring elderly patients from indoor or outdoor environments. The system uses a biosignal sensor worn by the patient and a Smartphone as a central node. The sensor data is collected and transmitted to the intelligent server through GPRS/UMTS to be analyzed. The prototype (UMHMSE) monitors the elderly mobility, location and vital signs such as Sp02 and Heart Rate. Remote users (family and medical personnel) might have a real time access to the collected information through a web application.
69 citations
TL;DR: This study study and evaluate the different methods for segmentation techniques and the main tendency of each algorithm with their applications, advantages and disadvantages is discussed.
Abstract: Evaluating the previous work is an important part of developing segmentation methods for the image analysis techniques. The aim of this paper is to give a review of digital image segmentation techniques. The problems of digital image segmentation represent great challenges for computer vision. The wide range of the problems of computer vision may make good use of image segmentation. This paper study and evaluate the different methods for segmentation techniques. We discuss the main tendency of each algorithm with their applications, advantages and disadvantages. This study is useful for determining the appropriate use of the image segmentation methods and for improving their accuracy and performance and also for the main objective, which designing new algorithms.
68 citations
TL;DR: In this paper, an integrated GPS-GSM system is proposed to track vehicles using Google Earth application, where the received GPS coordinates are filtered using a Kalman filter to enhance the accuracy of measured position.
Abstract: An integrated GPS-GSM system is proposed to track vehicles using Google Earth application. The remote module has a GPS mounted on the moving vehicle to identify its current position, and to be transferred by GSM with other parameters acquired by the automobile’s data port as an SMS to a recipient station. The received GPS coordinates are filtered using a Kalman filter to enhance the accuracy of measured position. After data processing, Google Earth application is used to view the current location and status of each vehicle. This goal of this system is to manage fleet, police automobiles distribution and car theft cautions.
63 citations
TL;DR: A new gesture recognition system based on image blocking is introduced and the gestures are recognized using the authors' suggested brightness factor matching algorithm and two different feature extraction techniques are applied.
Abstract: The rich information found in the human gestures makes it possible to be used for another language which is called the sign language, this kind of intuitive interface can be used with human-made machines/devices as well, we herein going to introduce a new gesture recognition system based on image blocking and the gestures are recognized using our suggested brightness factor matching algorithm, we have applied two different feature extraction techniques, the first one based on features extracted from edge information and the other one based on a new technique for centre of mass normalization based on block scaling instead of coordinates shifting; we have achieved 83.3% recognition accuracy in first technique with significant and satisfactory recognition time of 1.5 seconds per gesture, and 96.6 % recognition accuracy with recognition time less than one second by eliminating the use of edge detector which consumes time, this paper focuses on appearance based gestures.
60 citations
TL;DR: An approach of plant classification which is based on the characterization of texture properties is introduced and the combined classifier learning vector quantization gave a high performance which is a superior than other tested methods.
Abstract: paper introduces an approach of plant classification which is based on the characterization of texture properties. We used the combined classifier learning vector quantization. We randomly took out 30 blocks of each texture as a training set and another 30 blocks as a testing set. We found that the combined classifier method gave a high performance which is a superior than other tested methods. The experimental results indicated that our algorithm is applicable and its average correct recognition rate was 98.7%.
57 citations
TL;DR: It is eastliblished that back propagation neural network works successfully for the purpose of classification and the issue of improving the fitness (weight adjustment) of Back propagation algorithm is addressed.
Abstract: an Artificial Neural Network classifier is a nonparametric classifier. It does not need any priori knowledge regarding the statistical distribution of the class in a giver selected data Source. While, neural network can be trained to distinguish the criteria used to classify easily in a generalized manner that allows successful classification the newly arrived inputs not used during training. Through this paper it is eastliblished that back propagation neural network works successfully for the purpose of classification. Back propagation suffers from getting stuck into Local Minima. Weight optimization in Back propagation can be optimized using the Genetic Algorithm (GA). The back propagation algorithm is improved by invoking Genetic algorithm, to improve the overall performance of the classifier. The performance of a fitness algorithm using the approach suggested by us is a Hybrid System that is being analyzed in this paper. In this paper the issue of improving the fitness (weight adjustment) of Back propagation algorithm is addressed. Some of the Advantages of Hybrid algorithms are: convergence speed will be increased and the local minima problem can be overcome. The proposed Hybrid Algorithm is to perform learning as a back propagation and optimize weights using GA for classification.
TL;DR: A formal classification of attacks and vulnerabilities that affect current internet banking systems is presented along with two attacks which demonstrate the insecurity of such systems.
Abstract: A formal classification of attacks and vulnerabilities that affect current internet banking systems is presented along with two attacks which demonstrate the insecurity of such systems. Based ona thorough analysis of current security models, we propose a guidelines for designing secure internet banking systems which are not affected by the presented attacks and vulnerabilities .
TL;DR: In this article, a comparative experiment of four methods to identify plants using shape features was accomplished, and two approaches have never been used in plants identification yet, Zernike moments and Polar Fourier Transform (PFT), were incorporated.
Abstract: Shape is an important aspects in recognizing plants. Several approaches have been introduced to identify objects, including plants. Combination of geometric features such as aspect ratio, compactness, and dispersion, or moments such as moment invariants were usually used toidentify plants. In this research, a comparative experiment of 4 methods to identify plants using shape features was accomplished. Two approaches have never been used in plants identification yet, Zernike moments and Polar Fourier Transform (PFT), were incorporated. The experimental comparison was done on 52 kinds of plants with various shapes. The result, PFT gave best performance with 64% in accuracy and outperformed the other methods.
TL;DR: Experimental results demonstrate that the F-measure criterion improves up to 99% using feature selection and combination of AdaBoost and DT classifiers, which is highly comparable, and outperforms the previous reported F-measures.
Abstract: In the semantic web, ontology plays an important role to provide formal definitions of concepts and relationships. Therefore, communicating similar ontologies becomes essential to provide ontologies interpretability and extendibility. Thus, it is inevitable to have similar but not the same ontologies in a particular domain since there might be several definitions for a given concept. This paper presents a method to combine similarity measures of different categories without having ontology instances or any user feedback in regard with alignment of two given ontologies. To align different ontologies efficiently, K Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree (DT) and AdaBoost classifiers are investigated. Each classifier is optimized based on the lower cost and better classification rate. Experimental results demonstrate that the F-measure criterion improves up to 99% using feature selection and combination of AdaBoost and DT classifiers, which is highly comparable, and outperforms the previous reported F-measures.
TL;DR: The method was able to achieve an accuracy of 91.67% sensitivity and 84.17% specificity which is comparable to what has been reported using the state-of-the-art Computer-Aided Detection system.
Abstract: When researchers are studying the detection of breast tumors, there has been great attention to the modern textural features analysis of breast tissues on mammograms. Detecting of masses in digital mammogram based on second order statistics has not been investigated in depth. During this study, the breast cancer detection was based on second order statistics. The extraction of the textural features of the segmented region of interest (ROI) is done by using gray level co-occurrence matrices (GLCM) which is extracted from four spatial orientations; horizontal, left diagonal, vertical and right diagonal corresponding to (0 o , 45 o , 90 o and 135 o ) and two pixel distance for three different block size windows (8x8, 16x16 and 32x32) . The results show that the GLCM at 0 o , 45 o , 90 o and 135 o with a window size of 8X8 produces informative features to classify between masses and non-masses. Our method was able to achieve an accuracy of 91.67% sensitivity and 84.17% specificity which is comparable to what has been reported using the state-of-the-art Computer-Aided Detection system.
TL;DR: DM and NBA approaches for network intrusion detection are discussed and it is suggested that a combination of both approaches has the potential to detect intrusions in networks more effectively.
Abstract: Intrusion detection has become a critical component of network administration due to the vast number of attacks persistently threaten our computers. Traditional intrusion detection systems are limited and do not provide a complete solution for the problem. They search for potential malicious activities on network traffics; they sometimes succeed to find true security attacks and anomalies. However, in many cases, they fail to detect malicious behaviours (false negative) or they fire alarms when nothing wrong in the network (false positive). In addition, they require exhaustive manual processing and human expert interference. Applying Data Mining (DM) techniques on network traffic data is a promising solution that helps develop better intrusion detection systems. Moreover, Network Behaviour Analysis (NBA) is also an effective approach for intrusion detection. In this paper, we discuss DM and NBA approaches for network intrusion detection and suggest that a combination of both approaches has the potential to detect intrusions in networks more effectively.
TL;DR: Implementation of model for multi-criteria GDSS in which the simulation data is the mutated genes that can causecancer is proposed, which is a computer-based system that can be utilized in detecting human gene mutations that cause disease.
Abstract: Analysis of genes expression can be done with the investigation of a particular microarray data for the description of a gen. This is done to identify whatgenes that were active in the human body. Detection of gene mutation is an activity that can provide contribution in the medical field. Detection of mutated gene is needed to avoid the diseases caused by themsuch as cancer. The detection of gene mutations can be performed by utilizing computer-based system. Group Decision Support System (GDSS) is a computer-based system that can be utilized in detecting human gene mutations that cause disease. The ELECTRE method, which is a Multi-Attribute DecisionMaking, is a method in modeling multi-criteria GDSS. In this paper we propose implementation of model for multi-criteria GDSS in which the simulation data is the mutated genes that can causecancer
TL;DR: The present study aimed to do the performance analysis of several data mining classification techniques using three different machine learning tools over the healthcare datasets.
Abstract: Health care data includes patient centric data, their treatment data and resource management data. It is very massive and information rich. Valuable knowledge i.e. hidden relationships and trends in data can be discovered from the application of data mining techniques on healthcare data. Data mining techniques have been used in healthcare research and known to be effective. The present study aimed to do the performance analysis of several data mining classification techniques using three different machine learning tools over the healthcare datasets. In this study, different data mining classification techniques have been tested on four different healthcare datasets. The standards used are percentage of accuracy and error rate of every applied classification technique. The experiments are done using the 10 fold cross validation method. A suitable technique for a particular dataset is chosen based on highest classification accuracy and least error rate.
TL;DR: An algorithm for iris recognition and classification using a system based on Local Binary Pattern and histogram properties as a statistical approaches for feature extraction and Combined Learning Vector Quantization Classifier as Neural Network approach for classification, in order to build a hybrid model depends on both features.
Abstract: Iris recognition is considered as one of the best biometric methods used for human identification and verification, this is because of its unique features that differ from one person to another, and its importance in the security field. This paper proposes an algorithm for iris recognition and classification using a system based on Local Binary Pattern and histogram properties as a statistical approaches for feature extraction , and Combined Learning Vector Quantization Classifier as Neural Network approach for classification, in order to build a hybrid model depends on both features. The localization and segmentation techniques are presented using both Canny edge detection and Hough Circular Transform in order to isolate an iris from the whole eye image and for noise detection .Feature vectors results from LBP is applied to a Combined LVQ classifier with different classes to determine the minimum acceptable performance, and the result is based on majority voting among several LVQ classifier. Different iris datasets CASIA, MMU1, MMU2, and LEI with different extensions and size are presented. Since LBP is working on a grayscale level so colored iris images should be transformed into a grayscale level. The proposed system gives a high recognition rate 99.87 % on different iris datasets compared with other methods.
TL;DR: In this article, the authors proposed Elliptic Curve Cryptography (EC) as a vast field for public key cryptography (PKC) systems and proved that ECC has the same level of security with relatively small key sizes.
Abstract: Importance of Elliptic Curves in Cryptography was independently proposed by Neal Koblitz and Victor Miller in 1985.Since then, Elliptic curve cryptography or ECC has evolved as a vast field for public key cryptography (PKC) systems. In PKC system, we use separate keys to encode and decode the data. Since one of the keys is distributed publicly in PKC systems, the strength of security depends on large key size. The mathematical problems of prime factorization and discrete logarithm are previously used in PKC systems. ECC has proved to provide same level of security with relatively small key sizes. The research in the field of ECC is mostly focused on its implementation on application specific systems. Such systems have restricted resources like storage, processing speed and domain specific CPU architecture.
TL;DR: The aim of this paper is to familiarize the reader with the mathematics behind the application of wavelets for edge detection, which is made use of for spectrum sensing applications, and present an adaptive algorithm which chooses a suitable wavelet system by analyzing the nature of the spectrum.
Abstract: Spectrum scarcity is one of the major issues faced in Wireless communication technology. Efficient spectrum utilization is of utmost importance to alleviate the problem of interference and reduced data rates. Cognitive Radios adapt themselves according to the available spectrum and thereby enhance transmission and reception of data, without affecting adjacent band users. The pre-requisite for such an objective is the precise calculation of spectrum boundaries. Many methods have been suggested and revised from time to time. The Wavelet Edge Detection is one of the most widely used Spectrum Sensing techniques. This technique observes the spatial distribution of spectral data at multiple resolutions. The aim of this paper is to familiarize the reader with the mathematics behind the application of wavelets for edge detection, which is made use of for spectrum sensing applications. The several variants of this scheme which was originally formulated by Mallat et al are discussed and the inherent flaws or complexities are pointed out. The importance of choosing a suitable wavelet system is explained. We then proceed further and present an adaptive algorithm which chooses a suitable wavelet system by analyzing the nature of the spectrum. The slope of the Power Spectral Density is used as an index to distinguish between sharp and blunt peaks. Sparse spectra with conspicuous peaks utilize Haar wavelet system whereas dense spectra with subtle peaks use Gaussian Wavelet System. Multi-scale sums are used since they produce more accurate results than multi-scale products. The simulations are carried out in the FM frequency band (88-108 MHz) using MATLAB.
TL;DR: An automatic tool for identifying learning styles based on the Felder-Silverman learning style model in a learning environment using a social book marking website such as www.tagme1.com is presented.
Abstract: Personalized adaptive systems rely heavily on the learning style and the learner's behavior. Due to traditional teaching methods and high learner/teacher ratios, a teacher faces great obstacles in the classroom. In these methods, teachers deliver the content and learners just receive it. Moreover, teachers can’t cope with the individual differences among learners. This weakness may be attributed to various reasons such as the high number of learners accommodated in each classroom and the low teaching skills of the teacher himself/herself, Therefore, identifying learning styles is a critical step in understanding how to improve the learning process. This paper presented an automatic tool for identifying learning styles based on the Felder-Silverman learning style model in a learning environment using a social book marking website such as www.tagme1.com . The proposed tool used the learners’ behaviour while they are browsing / exploring their favorite web pages in order to gather hints about their learning styles. Then the learning styles were calculated based on the gathered indications from the learners' database. The results showed that the proposed tool recognition accuracy was 72% when we applied it on 25 learners with low number of links per learner. Recognition accuracy increased to 86.66% when we applied it on 15 learners with high number of links per learner.
TL;DR: In this paper, a Bangla parser based on the context free grammar is proposed. But the proposed parser is a predictive parser and the parse table for recognizing Bangla grammar is constructed using the left factoring.
Abstract: We describe a Context Free Grammar (CFG) for Bangla language and hence we propose a Bangla parser based on the grammar. Our approach is very much general to apply in Bangla Sentences and the method is well accepted for parsing a language of a grammar. The proposed parser is a predictive parser and we construct the parse table for recognizing Bangla grammar. Using the parse table we recognize syntactical mistakes of Bangla sentences when there is no entry for a terminal in the parse table. If a natural language can be successfully parsed then grammar checking from this language becomes possible. The proposed scheme is based on Top down parsing method and we have avoided the left recursion of the CFG using the idea of left factoring.
TL;DR: This paper presents the PHP framework for database management based on the MVC pattern, which is very useful for the architecture of web applications, separating the model, view and controller of a web application.
Abstract: PHP is a powerful language to develop dynamic and interactive web applications. One of the defining features of PHP is the ease for developers to connect and manipulate a database. PHP prepares the functions for database manipulation. However, database management is done by the Structure Query Language (SQL). Most novice programmers often have trouble with SQL syntax. In this paper, we present the PHP framework for database management based on the MVC pattern. The MVC pattern is very useful for the architecture of web applications, separating the model, view and controller of a web application. The PHP framework encapsulated, common database operations are INSERT, UPDATE, DELETE and SELECT. Developers will not be required to consider the specific SQL statement syntax, just to call it the method in the model module. In addition, we use White-Box testing for the code verification in the model module. Lastly, a web application example is shown to illustrate the process of the PHP framework.
TL;DR: The results show that the optimized fuzzy logic method for Magnetic Resonance Imaging (MRI) brain images segmentation effectively segmentedMRI brain images with spatial information, and the segmented MRI normal brain image and MRI brain image with tumor can be analyzed for diagnosis purpose.
Abstract: In this paper, an optimized fuzzy logic method for Magnetic Resonance Imaging (MRI) brain images segmentation is presented. The method is a technique based on a modified fuzzy c-means (FCM) clustering algorithm. The FCM algorithm that incorporates spatial information into the membership function is used for clustering, while a conventional FCM algorithm does not fully utilize the spatial information in the image. The advantages of the algorithm are that it is less sensitive to noise than other techniques, and it yields regions more homogeneous than those of other methods. Originality of this research is the methods applied on a normal MRI brain image and MRI brain images with tumor, and analyze the area of tumor from segmented images. The results show that the method effectively segmented MRI brain images with spatial information, and the segmented MRI normal brain image and MRI brain images with tumor can be analyzed for diagnosis purpose. In order to identify the area of abnormal mass of MRI brain images with tumor, it is resulted that the area is identified from 8.38 to 25.57 cm 2 .
TL;DR: This paper used one, two and three hidden layers and the modified additional momentum term to improve the character recognition capability of feed-forward back-propagation neural network with 182 English letters.
Abstract: This work is focused on improving the character recognition capability of feed-forward back-propagation neural network by using one, two and three hidden layers and the modified additional momentum term. 182 English letters were collected for this work and the equivalent binary matrix form of these characters was applied to the neural network as training patterns. While the network was getting trained, the connection weights were modified at each epoch of learning. For each training sample, the error surface was examined for minima by computing the gradient descent. We started the experiment by using one hidden layer and the number of hidden layers was increased up to three and it has been observed that accuracy of the network was increased with low mean square error but at the cost of training time. The recognition accuracy was improved further when modified additional momentum term was used.
TL;DR: In this article, one of the current leading video game companies was selected in order to perform an initial security assessment, which provided a starting point upon which specific goals and procedures were determined to help mitigate those risks.
Abstract: IT security issues are an important aspect for each and every organization within the video game industry. Within the video game industry alone, you might not normally think of security risks being an issue. But as we can and have seen in recent news, no company is immune to security risks no matter how big or how small. While each of these organizations will never be exactly the same as the next, there are common security issues that can and do affect each and every video game company. In order to properly address those security issues, one of the current leading video game companies was selected in order to perform an initial security assessment. This security assessment provided a starting point upon which specific goals and procedures were determined to help mitigate those risks. The information contained within was initially completed on the case study but has been generalized to allow the information to be easily applied to any video game company.
TL;DR: The findings suggest national culture significantly influence on participants’ attitude to sharing knowledge in IT-enabled virtual teams.
Abstract: Knowledge can be classified as explicit and implicit knowledge. This study investigates how national or regional culture influences on the explicit and implicit knowledge sharing behaviour in the context of multi-national virtual teams. The findings suggest national culture significantly influence on participants’ attitude to sharing knowledge in IT-enabled virtual teams. This study provides some suggestions for the virtual team management.
TL;DR: This paper presents the method developed to search and retrieve the similar image using bit plane image and uses simple Euclidean distance to compute the similarity measures of images for Content Based Image Retrieval application.
Abstract: Users needing is to store and index the image data and retrieved the image on feature vector derived by the user. Content Based Image Retrieval (CBIR) is search engine to retrieving the desired image automatically from the large image database having different categories. Retrieving the relevant images from the database by using feature vector is the challenging and important task. It is also need to retrieve the images from variety of the domain that is the application of CBIR that domains are medicine, crime prevention ,Biometrics, architecture, Fashion and publishing. This paper present the method developed to search and retrieve the similar image using bit plane image. Bit plane images are formed by using threshold and using bit plane slicing. Mean, standard deviation and third moment of row and column pixel distribution of bit plane image is used as a feature vector. We use simple Euclidean distance to compute the similarity measures of images for Content Based Image Retrieval application. The average precision and average recall of each image category and over all precision and recall is considered for the performance measure.
TL;DR: This paper presents a differential compression algorithm that is based on production of difference sequences according to op-code table in order to optimize the compression of homologous sequences in dataset to achieve up to 195-fold compression rate corresponding to 99.4% space saving.
Abstract: Modern biological science produces vast amounts of genomic sequence data. This is fuelling the need for efficient algorithms for sequence compression and analysis. Data compression and the associated techniques coming from information theory are often perceived as being of interest for data communication and storage. In recent years, a substantial effort has been made for the application of textual data compression techniques to various computational biology tasks, ranging from storage and indexing of large datasets to comparison of genomic databases. This paper presents a differential compression algorithm that is based on production of difference sequences according to op-code table in order to optimize the compression of homologous sequences in dataset. Therefore, the stored data are composed of reference sequence, the set of differences, and differences locations, instead of storing each sequence individually. This algorithm does not require a priori knowledge about the statistics of the sequence set. The algorithm was applied to three different datasets of genomic sequences, it achieved up to 195-fold compression rate corresponding to 99.4% space saving.