| Project/Area Number |
04452180
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
電子通信系統工学
|
|---|
| Research Institution | University of Tokyo |
|---|
Principal Investigator |
HATORI Mitsutoshi Univ.of Tokyo, Fuculty of Engineering Professor, 工学部, 教授 (60010790)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SAITO Takahiro Kanagawa Univ, Professor, 工学部, 教授 (10150749)
|
|---|
| Project Period (FY) |
1992 – 1993
|
| Project Status |
Completed (Fiscal Year 1993)
|
| Budget Amount *help |
¥5,800,000 (Direct Cost: ¥5,800,000)
Fiscal Year 1993: ¥1,000,000 (Direct Cost: ¥1,000,000)
Fiscal Year 1992: ¥4,800,000 (Direct Cost: ¥4,800,000)
|
|---|
| Keywords | wavelet transform / descrete biorthogonal wavelet transform / quantization noise analysis / image coding / adaptive edge detection / zero crossing representation / noise filtering / motion compensation interpolation / 適応的エッヂ検出 / 画像再構成 / 多重解像度解析 / 最適ビット配分 / 構造的特徴 / マルチスケール抽出 / フラクタル次元 / ゼロ交差情報 / 凸集合の反復射影法 |
|---|
| Research Abstract |
To provide foundations to techniques for dealing high definition and high resolution images, we conducted researches on wavelet transforms and its applications to image coding and processing.
1. Image Coding by using Wavelet Transform : We designed a novel biorthogonal wavelet transform, for efficient image coding, which has linear phase property and satisfies perfect reconstruction condition. We derived optimal bit-allocation through precise analysis of quantization errors. The proposed image coding scheme with the designed filter outperforms conventional subband coding methods.
2. Adaptive Edge Detection by using Wavelet Transform : In order to stably extract various edge features of images, we proposed a edge detection method based on wavelet transform. The method changes a parameter by estimated fractal dimension.
3. Image Reconstruction from Zero Crossing Representation : We investigated image reconstruction from zero-crossing representation obtained by wavelet transform. Our zero-crossing representation makes use of maximum points in addition to the zero-crossing points. We derived an iterative method that stably reconstructs original image from its zero-crossing representation, and also found a condition for filters to be robust to reconstruction errors.
4. Image Sequence Processing by using Wavelet Transform : We investigated segmentation of image sequences by using wavelet transform and applied the segmentation scheme to motion compensated interpolation.
|
