A Two-dimensional Direct Combined Model for Facial Hallucination | |
---|---|
學年 | 101 |
學期 | 2 |
發表日期 | 2013-05-20 |
作品名稱 | A Two-dimensional Direct Combined Model for Facial Hallucination |
作品名稱(其他語言) | |
著者 | Tu, Ching-Ting; Lin, Kung-Hung; Liu, Tyng-Luh |
作品所屬單位 | 淡江大學資訊工程學系 |
出版者 | |
會議名稱 | MVA 2013 IAPR International Conference on Machine Vision Applications |
會議地點 | Kyoto, Japan |
摘要 | This study developed a face hallucination system based on a novel two-dimensional direct combined model (2DDCM) approach that employs a large collection of low-resolution/high-resolution facial pairwise training examples. The proposed 2DDCM approach achieves face hallucination by addressing three key issues. First, we directly combine each low-resolution and high-resolution pairwise image in a concatenated form in order to completely preserve their relationship. Second, images are formed as two–dimensional matrices instead of vectors in order to preserve the facial geometry. Third, both the vertical and the horizontal facial-geometry features are considered in 2DDCM approach. Experiments demonstrate our approach can synthesize high quality reconstructed facial images from given low-resolution images. |
關鍵字 | |
語言 | en |
收錄於 | |
會議性質 | 國際 |
校內研討會地點 | |
研討會時間 | 20130520~20130523 |
通訊作者 | |
國別 | JPN |
公開徵稿 | |
出版型式 | |
出處 | |
相關連結 |
機構典藏連結 ( http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/100476 ) |