使用手機連接雲端視訊服務平台的研究

Abstract

本論文主要地目的是研究出如何將現今具有普遍性的手機連結到雲端視訊服務平台上，並幫助那些平常沒有電腦(例如：偏遠地方的使用者)或者無法隨身攜帶電腦的使用者使用手機也可以獲得基本的服務平台服務。 在此我們提出了基於中文斷詞技術PAT-Tree做關鍵詞擷取，並實作出本實驗室研發的通用型機率決策神經網路之紋理辨識並結合K-means clustering演算法做影像檢索，其中也提出了K值演算法去決定K-means clustering所需的K值。

The purpose of this thesis focuses on how to use mobile phones to connect the cloud computing video service platform. Then, we can help those who do not have a computer(e.g. The people in the remote area) or carry a computer all the time to access to basic services by using mobile phones. Now, this thesis purpose to keyword extraction based on Chinese Word Segmentation Technology - PAT-Tree, and implemented “Texture Recognition By Generalizes Probabilistic Decision-Based Neural Networks with K-means clustering to image retrieval. This thesis also purpose K-value algorithm to determine the value of K for K-means clustering.