The problem inherent to any digital image is the large amount of bandwidth required for transmission or storage. This has driven the research area of image compression to develop algorithm that compress images to lower data rates with better quality. This research present, a new approach to image compression based on clustering. This new approach includes new objective function, and its minimization by energy function based on unsupervised two dimensional fuzzy Hopfield neural network. New objective function consists of a combination of classification entropy function and average distance between image pixels and cluster centers. After applying new method on gray scale sample images at different number of clusters, better compression ratio and signal to noise ratio was observed. The new method is also a new clustering analysis method, and it provides more compact and separate clustering.
ion:The work explores the potentiality of a clonal selection algorithm and it's hybridizing with the genetic algorithm GA in cursive and discrete handwritten English character recognition. In particular, a retraining scheme for the clonal selection algorithm is formulated for better recognition rates. Empirical study with a dataset (which contains about 100 handwritten samples for 26 characters taken from 30 persons) shows that the proposed approach exhibits very good generalization ability, such that results reported recognition accuracy reached to 100% for the recognition of characters that have been used in building database, and an average recognition accuracy of about 94% for other characters.
اﻟﻣﻠﺧص ّ :اﻟﻧﺳﯾﻠﻲ اﻻﺧﺗﯾﺎر ارزﻣﯾﺔ ﺧو اﺳﺗﺧدام ﻗﺎﺑﻠﯾﺔ اﻟﻌﻣل ﯾﻌرض clonal selection اﻟﻣﻧﺎﻋﯾﺔ اﻟﺟﯾﻧﯾﺔ ارزﻣﯾﺔ اﻟﺧو ﻣﻊ وﺗﻬﺟﯾﻧﻬﺎ GA ﻓ ﺑﺧط اﻟﻣﻛﺗوﺑﺔ اﻟﻣﺗﺻﻠﺔ ﯾﺔ اﻻﻧﻛﻠﯾز اﻟﻠﻐﺔ أﺣرف ﺗﻣﯾﯾز ﻲ اﻟﯾد . اﻟﻧﺳﯾﻠﻲ اﻻﺧﺗﯾﺎر ارزﻣﯾﺔ ﻟﺧو ﺗﯾب اﻟﺗر ﻣﻌﺎد ﻫﯾﻛل اﻟﺗﺣدﯾد وﺟﻪ ﻋﻠﻰ اﺳﺗﺧدم ﻓﻘد clonal selection ﻣﻣﻛﻧﺔ ﺗﻣﯾﯾز ﻧﺳﺑﺔ أﻓﺿل ﻟﺗﺣﻘﯾق . ﻣﺟﻣوﻋﺔ ﻋﻠﻰ ﯾت أﺟر ﯾﺑﯾﺔ ﺗﺟر اﺳﺔ در ﺧﻼل ﻓﻣن ﻣن ﻣﻛوﻧﺔ 100 ﻧﻣوذج sample اﻻﻧ اﻷﺣرف ﻣن ﻣن ﻷﻛﺛر اﻟﯾد ﺑﺧط اﻟﻣﻛﺗوﺑﺔ ﯾﺔ ﻛﻠﯾز 30 ﺷﺧﺻ ﺎ ً إﻟﻰ ﺗﺻل ﺗﻣﯾﯾز ﻧﺳﺑﺔ ﺗﺣﻘﯾق ﻓﻲ اﻟﻌﺎﻟﯾﺔ ﻓﻌﺎﻟﯾﺗﻬﺎ ارزﻣﯾﺔ اﻟﺧو ﻫذﻩ أﺛﺑﺗت 100 % ﺑﺎﻟﻧﺳﺑﺔ
The aim of this work is to recognize the printed Latin's characters. In this work two methods for constructing the feature space are used. These methods are Variance and Fractal dimension methods, as a result they have real values for every character in the Latin's language, and from these values they constructed the feature space extractions for every character in the Latin's language. After that, these features are given to the Back Propagation network for recognizing the characters.The result is a highest recognition for the characters is obtained, it is about 82.75% characters while the unrecognized characters are 17.25.
In this research, an intelligent computer system is designed for recognizing printed Russian letters by extracting features of the letter by finding the Eigen values which then used for training and testing the artificial neural network used in this work namely, Elman NN. This network is used as a tool for decision making. Data is entered using a flatbed scanner which results in high extensity, fineness and homogeneous BMP extension images. The programs are implemented by Matlab language, the software include image enhancement techniques, image segmentation, resize the segmented image and features extraction dependent on Eigen values .These values are then used to train and test the Elman Neural Network. In this work the pass ratio of recognition up to 90 % .
In this research, a new method is discovered (combined method) to accelerate the backpropagation network by using the expected values of source units for updating weights, we mean the expected value of unit by the sum of the output of the unit and its error term multiplied by the factor Beta to accelerate the algorithm and also adjust the value of learning coefficient continuously if the value of energy function E decreases the learning rate is increased by a factor , if the value of the energy function E increases , the value of the learning rate is decreased by a factor . To obtain the optimal weight with minimum iteration and minimum time, we applied a new method on many applications to prove the result of this method (pattern compression, encoding and recognition on Arabic, English digits and alphabetic.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.