日本語フィールド
著者：Hiroki Matsunaga, Hajime Omura, Ryuji Ohura, Teruya Minamoto題名：Daubechies wavelet-based method for early esophageal cancer detection from flexible spectral imaging color enhancement image発表情報：Advances in Intelligent Systems and Computing 巻: 448 ページ: 939-948キーワード：概要：抄録：We propose a new method for detecting early esophageal cancer from Flexible spectral Imaging Color Enhancement (FICE) mode images based on the Daubechies wavelet transform. In our method, we convert an image into CIEL*a*b* color space and use the a* components. Next, we divide the a* components into small blocks, apply two types of Daubechies wavelet transforms to each block, and then obtain the low- and high-frequency components at each block. The histogram of the low-frequency components tends to be positioned at the right and the left of a particular value for abnormal and normal regions, respectively. The histogram of the high-frequency components for an abnormal region has a longer tail. We describe the detection procedure of the abnormal regions in detail and present experimental results demonstrating that our method is able to detect early esophageal cancer from FICE images based on these features.英語フィールド
Author：Hiroki Matsunaga, Hajime Omura, Ryuji Ohura, Teruya MinamotoTitle：Daubechies wavelet-based method for early esophageal cancer detection from flexible spectral imaging color enhancement imageAnnouncement information:Advances in Intelligent Systems and Computing Vol: 448 Page: 939-948An abstract：We propose a new method for detecting early esophageal cancer from Flexible spectral Imaging Color Enhancement (FICE) mode images based on the Daubechies wavelet transform. In our method, we convert an image into CIEL*a*b* color space and use the a* components. Next, we divide the a* components into small blocks, apply two types of Daubechies wavelet transforms to each block, and then obtain the low- and high-frequency components at each block. The histogram of the low-frequency components tends to be positioned at the right and the left of a particular value for abnormal and normal regions, respectively. The histogram of the high-frequency components for an abnormal region has a longer tail. We describe the detection procedure of the abnormal regions in detail and present experimental results demonstrating that our method is able to detect early esophageal cancer from FICE images based on these features.