Image segmentation using a combined DCT and neural networks technique
碩士 === 立德管理學院 === 應用資訊研究所 === 93 === Almost all digital images are stored in compressed formats which are defined by joint picture expert group (JPEG) and are widely adopted on image database. The fast image extraction algorithm is applied and the image is directly extracted Discrete Cosine Transfor...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2005
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/01765362348595012247 |
| Summary: | 碩士 === 立德管理學院 === 應用資訊研究所 === 93 === Almost all digital images are stored in compressed formats which are defined by joint picture expert group (JPEG) and are widely adopted on image database. The fast image extraction algorithm is applied and the image is directly extracted Discrete Cosine Transform (DCT) coefficients. With these DCT coefficients, the original image is transformed to a new image for the proposed automatic segmentation technique.
Based on the histogram analysis, the new thresholding method is proposed and is combined with neural networks technique. The experiment results display that the computation of automatic segmentation is more efficient and the quality achieves the human vision as well.
|
|---|