تصنيف المساجد الجامعة بطريقة تمييز النمط

Abstract EN

Congregational Mosques Classification is an important issue for Islamic Architecture studies which introduced many different classifications.

This paper aims to introduce a new classification of the congregational mosques using the pattern recognition digital technique.

The cases selected for the study covers most of the period and space of the Islamic for a more objective and accurate classification.

The paper takes up an unsupervised texture image classification algorithm which uses a competitive neural network.

The classification is done by assigning data to one of the fixed number of possible classes then classifying image segments into a given number of classes using segments features by using Kohonen competitive neural network and through a process of self-organization, configures the output units, representing the selected congregational mosques, into a spatial map including the resulted classification types.

The conclusions support some known statements and contradict others within Islamic Architecture Theory, so certain modifications are needed according to the concluded statements after corroborating them in future studies.