NATIONAL YUNLIN UNIVERSITY OF SCIENCE AND TECHNOLOGY
发明人:
CHANG, HSUAN-TING,张轩庭,張軒庭,LIN, KUN-PEI,林昆霈
申请号：
TW107110708
公开号：
TW201942868A
申请日:
2018.03.28
申请国别(地区):
TW
年份:
2019
代理人:
摘要：
A method for identifying image containing tissue symptoms. The method includes dividing an image into blocks； capturing a plurality of feature vectors of each of the blocks within the image, where said feature vector is determined based on a statistics of pixel value variation within the block； generating a set of coefficients of variation based on a statistics of the feature vectors of all blocks within the image； determining whether the image being a symptom image based on the set of coefficients of variation of the image and a first classifying model； and determining whether the block of the image being a symptom block based on the feature vectors of the block and a second classifying model.一種用於辨識組織病徵影像的方法，包含：將一影像切割為複數個區塊；擷取該影像的每一區塊的複數個特徵向量，所述特徵向量是基於該區塊中像素值的變化統計所決定；基於該影像的所有區塊的該等特徵向量的一統計而產生代表該影像的一組變異係數；基於該影像的該組變異係數及一第一分類模型決定該影像是否為一病徵影像；基於該影像的該區塊的該等特徵向量及一第二分類模型決定該影像的該區塊是否為一病徵區塊。
