Keywords : Segmentation


Medical Image Classification Using Different Machine Learning Algorithms

Sami H. Ismael; Shahab W. Kareem; Firas H. Almukhtar

AL-Rafidain Journal of Computer Sciences and Mathematics, 2020, Volume 14, Issue 1, Pages 135-147
DOI: 10.33899/csmj.2020.164682

The different types of white blood cells equips us an important data for diagnosing and identifying of many diseases. The automation of this task can save time and avoid errors in the identification process. In this paper, we explore whether using shape features of nucleus is sufficient to classify white blood cells or not. According to this, an automatic system is implemented that is able to identify and analyze White Blood Cells (WBCs) into five categories (Basophil, Eosinophil, Lymphocyte, Monocyte, and Neutrophil). Four steps are required for such a system; the first step represents the segmentation of the cell images and the second step involves the scanning of each segmented image to prepare its dataset. Extracting the shapes and textures from scanned image are performed in the third step. Finally, different machine learning algorithms such as (K* classifier, Additive Regression, Bagging, Input Mapped Classifier, or Decision Table) is separately applied to the extracted (shapes and textures) to obtain the results. Each algorithm results are compared to select the best one according to different criteria’s.
 

Information Hiding Based on Chan-Vese Algorithm

Samia Sh. Lazar; Nadia M. Mohammed

AL-Rafidain Journal of Computer Sciences and Mathematics, 2011, Volume 8, Issue 2, Pages 99-110
DOI: 10.33899/csmj.2011.7876

The process of data transfer via the Internet becomes an easy process as a result of the great advances in networking technologies, and now many people can communicate with each other easily and quickly through them.Because the online environment is general and open, the unauthorized one can control information were transmitted between any two parts and interception of getting access for it, because of that there is an emergency need for write covered, which is the science of hiding secret information in a digital cover such as an images, so it is impossible for the normal person and others unauthorized to detected or perceives. In this paper, the technology in the field of information hiding in the images is developed, where first, the cover (PNG, BMP) image is segmented using Chan-Vese algorithm, then the text will hide in the segmented image depending on the areas of clipping.The standards (PSNR, BER) are used to measure technical efficiency. In addition the algorithm of this technique is implemented in Matlab.

Segmentation of Brain Tumor Images Using Genetic Algorithms

Mohameed Nathem; Manar Y. Kashmola; Dhuha Basheer Abdullah

AL-Rafidain Journal of Computer Sciences and Mathematics, 2009, Volume 6, Issue 1, Pages 31-45
DOI: 10.33899/csmj.2009.163783

In this research, the brain images resulting from computerized tomography  (CT) have been used in order to determine the tumor area in the brain. the steps started by preprocessing operation to the image before inputting it to algorithm .the image was converted to binary image in order to segment the image, later on into equal segments ,then the correlation coefficient  was found among these segments ,these values used as fitness function in the genetic algorithm in order to different iate between segments ,the result of the genetic algorithm was segment numbers which will be merged to form the sub-images ,then continuing these steps till determining the tumor approximate location. Another approach of image segmentation has been used without using the genetic algorithm by choosing the segments though a certain condition not randomly. Satisfying results have been reached in both approaches, but in different execution times. In both approaches tumor location was determined approximately, as a result the genetic algorithm was succeeded in about 85% while the first algorithm determine 80% of tumor locations.
 

Segmentation of Arabic Word into Letters and Recognition

Faten Basher Abd Alahad; Enaam Ghanim Saeed

AL-Rafidain Journal of Computer Sciences and Mathematics, 2005, Volume 2, Issue 2, Pages 123-141
DOI: 10.33899/csmj.2005.164092

In this research a system for Arabic word segmentation into letters and recognition has been designed by dividing it into five groups (plosives, fricatives, nasals, glide and semi-vowel sounds) based on articulatory phonetics. This system include four main stages:
Stage one: Endpoint algorithm has been used to identify the  beginning and the end of word.
Stage two: A new segmentation algorithm has been suggested and implemented depending upon time domain features and Arabic phonology rules.
Stage three: This stage includes letters feature extraction depending on linear predictive coding and system data base constructing which include vectors of features for the segmented letters from words and regarding letter recurring in different positions in the word.
Stage four: Includes Arabic letters recognition according to articulation which entails using Dynamic Time Warping (DTW) method that uses dynamic programming basics to obtain the matching path for the least distance accumulated value, where the word used in segmentation and recognition belongs to the four persons who create the data base and the results were in consistence which ranged from (75- 80)%