図形認識における特徴抽出の一方法

概要

論文の詳細を見る
In general,feature extraction in the picture recognition requires formidable computation.So,we developed a feature extraction method which incorporates local data operation and relatively simple algorithm.This method,named Georletric Factor Method,utilizes the geometry computation theory in Perceptrons.The pictures are analyzed and feature values ln the Picture are extracted according to the geometric factors as dehned in the method.In a simulation based on the method,feature values of randomly generated circles,triangles and rectangles are extracted and evaluated.Also,the method is tested on classincation of human white cells for a medical picture processing application.
一般社団法人情報処理学会の論文
1975-05-15