國立高雄大學圖資館 |

Language: English

Back

基於增長層級式SOM之自動影像註解方法 = Automatic Imag...

國立高雄大學資訊管理學系碩士班

基於增長層級式SOM之自動影像註解方法 = Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps

Record Type:	Language materials, printed : monographic
Paralel Title:	Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps
Author:	莊智翔,
Secondary Intellectual Responsibility:	國立高雄大學
Place of Publication:	[高雄市]
Published:	撰者;
Year of Publication:	2009[民98]
Description:	50面圖,表 : 30公分;
Subject:	Automatic Image Annotation
Subject:	增長層級式自我組織圖
Online resource:	http://handle.ncl.edu.tw/11296/ndltd/18573884670521935511
Notes:	參考書目：面
Notes:	指導教授：楊新章
Summary:	隨著數位影像科技發展，數位影像資料量快速成長，自動影像檢索技術的發展與應用逐漸成為重要之研究課題。為能有效檢索影像，影像語意之淬取與表達為重要之步驟。傳統上影像語意之淬取與表達大致上分為直接以影像視覺特徵表達與以文字註解表達兩類。然而使用視覺特徵具有不易淬取與語意混淆之缺點。反之，文字註解之語意明確，且容易進行檢索。但對大量影像進行註解為一十分花費時間與人力之工作。本研究重點在於發展一適用於影像文件上之自動影像註解方法。在本研究中將採用增長層級式自我組織圖(The Growing Hierarchical Self-Organizing Map，GHSOM)來協助發掘影像文件與註解文件之間隱含之關聯，並使用此關聯進行影像註解。我們將先使用GHSOM分別對影像與註解進行分群與階層產生，並提出一階層比對方法以發掘影像群集與註解群集間之關聯。新進之影像即可利用此關聯進行註解。本文以一參考資料庫進行實驗，實驗結果證明本方法在進行自動影像註解上頗為可行。 With the improve of digital image technology, digital image data was increasing, the issue that research about automatic image annotation technology and applications become more and more important. In order to image retrieval efficiently, how to extracting and present the semantic of image was an important process in image retrieval. Traditional image semantic extraction and present usually can divide to image vision and text annotation. But the vision feature has a disadvantage that they are hard to extract and semantic confused. On the other hand, the image semantic can presented clear by text annotation, and easily to retrieval. But it is usually time-consuming and manpower. Our purpose are developing an automatic image annotation method for image data, we will use Growing Hierarchical Self-Organizing Map (GHSOM) to help us discovery the concealed relations between image data and annotation data, and finish image annotation. We will use GHSOM to cluster for image and annotation data, generate each hierarchical structure separately, and then, we propose a hierarchical mapping method to discover the relations between image cluster and annotation cluster. New image can be annotated by that. We use an image dataset to experiment and have a good result to prove our method for automatic image annotation is practicable.

基於增長層級式SOM之自動影像註解方法 = Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps
莊, 智翔

基於增長層級式SOM之自動影像註解方法 = Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps / 莊智翔撰 - [高雄市] : 撰者, 2009[民98]. - 50面 ; 圖,表 ; 30公分.
參考書目：面指導教授：楊新章.
Automatic Image Annotation增長層級式自我組織圖

基於增長層級式SOM之自動影像註解方法 = Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps
LDR:03626nam a2200277 4500 001 220402
005 20170214095234.0
009 220402
010 0 $b 平裝
010 0 $b 精裝
100 $a 20170214y2009 k y0chiy09 ea
101 1 $a chi $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 基於增長層級式SOM之自動影像註解方法 $d Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps $z eng $f 莊智翔撰
210 $a [高雄市] $c 撰者 $d 2009[民98]
215 0 $a 50面 $c 圖,表 $d 30公分
300 $a 參考書目：面
300 $a 指導教授：楊新章
328 $a 碩士論文--國立高雄大學資訊管理學系碩士班
330 $a 隨著數位影像科技發展，數位影像資料量快速成長，自動影像檢索技術的發展與應用逐漸成為重要之研究課題。為能有效檢索影像，影像語意之淬取與表達為重要之步驟。傳統上影像語意之淬取與表達大致上分為直接以影像視覺特徵表達與以文字註解表達兩類。然而使用視覺特徵具有不易淬取與語意混淆之缺點。反之，文字註解之語意明確，且容易進行檢索。但對大量影像進行註解為一十分花費時間與人力之工作。本研究重點在於發展一適用於影像文件上之自動影像註解方法。在本研究中將採用增長層級式自我組織圖(The Growing Hierarchical Self-Organizing Map，GHSOM)來協助發掘影像文件與註解文件之間隱含之關聯，並使用此關聯進行影像註解。我們將先使用GHSOM分別對影像與註解進行分群與階層產生，並提出一階層比對方法以發掘影像群集與註解群集間之關聯。新進之影像即可利用此關聯進行註解。本文以一參考資料庫進行實驗，實驗結果證明本方法在進行自動影像註解上頗為可行。 With the improve of digital image technology, digital image data was increasing, the issue that research about automatic image annotation technology and applications become more and more important. In order to image retrieval efficiently, how to extracting and present the semantic of image was an important process in image retrieval. Traditional image semantic extraction and present usually can divide to image vision and text annotation. But the vision feature has a disadvantage that they are hard to extract and semantic confused. On the other hand, the image semantic can presented clear by text annotation, and easily to retrieval. But it is usually time-consuming and manpower. Our purpose are developing an automatic image annotation method for image data, we will use Growing Hierarchical Self-Organizing Map (GHSOM) to help us discovery the concealed relations between image data and annotation data, and finish image annotation. We will use GHSOM to cluster for image and annotation data, generate each hierarchical structure separately, and then, we propose a hierarchical mapping method to discover the relations between image cluster and annotation cluster. New image can be annotated by that. We use an image dataset to experiment and have a good result to prove our method for automatic image annotation is practicable.
510 1 $a Automatic Image Annotation Based on Growing Hierarchical Self-Organizing Maps $z eng
610 1 $a Automatic Image Annotation $a Neural Network $a Text
610 0 $a 增長層級式自我組織圖 $a 文件分群 $a 自動影像註解 $a 類神經網路
681 $a 008M/0019 $b 464105 4488 $v 2007年版
700 1 $a 莊 $b 智翔 $4 撰 $3 354066
712 0 2 $a 國立高雄大學 $b 資訊管理學系碩士班 $3 353936
801 0 $a tw $b 國立高雄大學 $c 20091020 $g CCR
856 7 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/18573884670521935511