المستودع الرقمي العراقي
Iraqi Digital Repository
الشبكات المتعلمة الرقمية متعددة الاصناف المحورة لاغراض فحص كريات الدم الحمراء == Modified Multi - Category Digital Learning Network For Red Blood Cell Inspection
اسم المؤلف:
سهاد قاسم غلام حسين حداد
اسم المشرف:
محمود حمزة المفرجي
الموضوع العام:
هندسة الكهرباء والالكترون والاتصالات
السنة:
2014
الموضوع الدقيق:
هندسة الالكترونيك
الدرجة:
ماجستير
الجامعة:
الجامعة التكنولوجية - قسم الهندسة الكهربائية
اللغة:
الانكليزية
مكان الجامعة:
بغداد
الكلمات الدلالية:
- Pattern Recognition
- RAM - based Weightless Neural Network
الصفحات الاولى:
📎 34T501 - p.pdf
المستخلص:
A pattern recognition system based on the n - tuple technique is developed and evaluated for use in classifying non - deterministic data with particular reference to medical image. The pattern recognition system presented in this work fulfills the requirements of simplicity and efficiency making it attractive to practical use in present day for industrial and medical environments. It is an effective solution for providing healthcare with reduced cost, especially for the rural areas and far away patients. Ordinary doctors (not specialist in blood diseases), will be able to perform extra - ordinary tasks.In this work Digital Learning Network has been designed for classification of different shapes of abnormal Red Blood Cells. Digital Learning Network is of low cost hardware and implementation, and one shot learning, using networks of RAMs. Many parameters have been investigated in details which affect the recognition rate. These parameters are presented to allow the system to be optimized, giving an increase in the performance of the system. Modification method of Feedback Digital Learning Network, which is an improving process of Digital Learning Network, has been implemented. The obtained results show that high performance (96.6%) can be achieved, providing evidence of the validity of the proposed technique.
مشاركة البيانات
Share Data