Text Detection based on Adaptive Edge Detection and its Application for Transparent Text Extraction
スポンサーリンク
概要
- 論文の詳細を見る
We propose methods for detecting and extracting transparent texts in natural scenes. First, we developed an adaptive edge detection method for finding connected components, even in low contrast images, which is common among transparent texts. Then the components are analyzed using the designed text-specific features, followed by character grouping into lines and word separation. We also developed a method for extracting transparent text, while preserving image contents, once their locations are detected. Experiments show the effectiveness of our adaptive edge detection method and our transparent text extraction.
- 一般社団法人電子情報通信学会の論文
- 2013-03-07
著者
関連論文
- Text Detection based on Adaptive Edge Detection and its Application for Transparent Text Extraction
- 学習型テキスト画像検出(テーマセッション,文字・文書の認識と理解及びアルゴリズム・システム評価)