Open AccessResearch on Unsupervised Computer Deep Learning Method for Color Image Recognition
Author(s) -
Shuqiang Du
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1915/4/042058
Subject(s) - artificial intelligence , computer science , pattern recognition (psychology) , deep learning , field (mathematics) , artificial neural network , unsupervised learning , feature (linguistics) , construct (python library) , focus (optics) , image (mathematics) , autoencoder , color image , computer vision , image processing , mathematics , philosophy , physics , pure mathematics , optics , programming language , linguistics
With the development of science and technology, face recognition has gradually become the focus of identity verification, which leads to the research of color image recognition. Compared with the traditional gray image, color image contains more information, which requires us to make full use of the complementary information between each color component to remove redundant information. Therefore, color image recognition technology has been an important research field in the field of vision. Unsupervised feature learning (hereinafter referred to as UFL) is an algorithm technology to automatically extract data hidden features, which is an important feature method combined with Deep learning (hereinafter referred to as DL) algorithm. Through the fusion of UFL and DL algorithm model, we can recognize and analyze color images. In this method, we will use self encoder, which can be used as UFL model. By stacking multiple self encoders, we can construct a deep neural network. Then, we can train the network of each hidden layer through unsupervised training, which will optimize the neural network supervised. Therefore, unsupervised learning method has been applied to CNN, which has become an important method in color image recognition.