A Hybrid of Fuzzy Integral and Convolution Neural Networks for Facial Image Classification

碩士 === 國立勤益科技大學 === 資訊工程系 === 106 === This paper proposed Evolutional Fuzzy Integral to promote the accuracy by combining several Convolution Neural Networks. And solve the problem of how to choice witch modle or optimization to use when you’re facing a classification issue. the input of Evolutional...

Full description

Bibliographic Details
Main Authors: Wei-Yu Tseng, 曾韋諭
Other Authors: Cheng-Jian Lin
Format: Others
Language:zh-TW
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/94p3tx
Description
Summary:碩士 === 國立勤益科技大學 === 資訊工程系 === 106 === This paper proposed Evolutional Fuzzy Integral to promote the accuracy by combining several Convolution Neural Networks. And solve the problem of how to choice witch modle or optimization to use when you’re facing a classification issue. the input of Evolutional Fuzzy Integral is the output of Convolution Neural Networks. The result will be compute by the rule of Sugeno or Choqute. Evolutional Fuzzy Integral can analysis the advantage of Convolution Neural Networks. And based on it to compute the output. So it can achieve purpose of Combining the each Convolution Neural Network according to the past experience. Which means it can automatically evaluate the advantage of each Convolution Neural Network.