METHOD FOR TRAINING THE NEURAL NETWORK METHOD FOR RECOGNING USING NEURAL NETWORK AND APPARATUS THEREOF

Disclosed are a learning method of a neural network, a recognizing method using the neural network, and an apparatus thereof. The recognizing method using the neural network includes the steps of: obtaining a feature vector generated from a hidden layer of the neural network according as input data...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: KANG, HYO A
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Disclosed are a learning method of a neural network, a recognizing method using the neural network, and an apparatus thereof. The recognizing method using the neural network includes the steps of: obtaining a feature vector generated from a hidden layer of the neural network according as input data to be recognized is inputted to an input layer of the neural network; and determining the reliability of a recognition result for the input data by using the feature vector and a plurality of clusters. Accordingly, the present invention can effectively prevent a recognition result with low reliability from being used. 뉴럴 네트워크의 학습 방법 및 뉴럴 네트워크를 이용한 인식 방법, 및 그 장치가 개시된다. 개시된 뉴럴 네트워크를 이용한 인식 방법은 인식하고자 하는 입력 데이터가 뉴럴 네트워크의 입력 레이어에 입력됨에 따라 뉴럴 네트워크의 히든 레이어로부터 생성된 특징 벡터를 획득하고, 특징 벡터 및 복수의 클러스터들을 이용하여, 입력 데이터에 대한 인식 결과의 신뢰도를 결정한다.