IMAGE RECOGNITION DEVICE CAPABLE OF CHANGING ARRANGEMENT AND COMBINATION OF WINDOWS USED IN IMAGE RECOGNITION IN ACCORDANCE WITH CONFIGURATION INFORMATION

PROBLEM TO BE SOLVED: To provide an image recognition device which prevents the occurrence of reduction of false recognition even when there is a pattern resembling a portion of a plurality of components constituting a recognition object.SOLUTION: A configuration information setting part 12 sets the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: YAMANO JUNJI, ONOE TAKAO
Format: Patent
Sprache:eng ; jpn
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:PROBLEM TO BE SOLVED: To provide an image recognition device which prevents the occurrence of reduction of false recognition even when there is a pattern resembling a portion of a plurality of components constituting a recognition object.SOLUTION: A configuration information setting part 12 sets the shape of a window for scanning in accordance with configuration information so as to include each of a plurality of components constituting a recognition object. Image feature amount calculation parts 7i1, i2... (M) calculate the feature amount for each shift position when the window is arranged at each of the plurality of shift positions. Feature amount identification parts 9i1, i2... (N) determine whether or not the components of the recognition object are included in a local area at each shift position. When the local area including each component is discovered, an object recognition part 11 performs comprehensive determination on whether or not the recognition object exists in accordance with the overall positional relationship between the plurality of local areas.SELECTED DRAWING: Figure 7 【課題】認識対象を構成する複数の構成要素のうち、一部のものと似通った絵柄が存在するとしても、誤認識の低下を発生させることがない画像認識装置を提供する。【解決手段】構成情報設定部12は、認識対象を構成する複数の構成要素のそれぞれが包含されるよう、構成情報に従って、スキャンのためのウィンドウの形状を設定する。画像特徴量計算部7i1、i2・・・(M個)は、複数のシフト位置のそれぞれにウィンドウが配置された際、各シフト位置について特徴量を算出する。特徴量識別部9i1、i2・・・(N個)は、各シフト位置の局所領域に、認識対象の構成要素が包含されているかどうかを判定する。各構成要素を包含した局所領域が発見された場合、客体認識部11は、それら複数の局所領域の全体的な位置関係に応じて、認識対象が存在するかどうかの総合判断を行う。【選択図】図7