现代电子技术
現代電子技術
현대전자기술
MODERN ELECTRONICS TECHNIQUE
2013年
17期
118-121
,共4页
尿沉渣%图像分割%多信息互补%SVM
尿沉渣%圖像分割%多信息互補%SVM
뇨침사%도상분할%다신식호보%SVM
urine sediment%image segmentation%multi-information complementary%SVM
针对尿沉渣图像中的红白细胞,提出了一种基于组合思想的分割方法,即对图像进行三层处理,将各层的分割结果进行融合,从而通过多信息互补的方法得到完整的分割结果。设计了两级集成SVM分类器对红白细胞进行识别。实验证明,提出的整套算法简洁高效,精度高,具有较强的普适性。
針對尿沉渣圖像中的紅白細胞,提齣瞭一種基于組閤思想的分割方法,即對圖像進行三層處理,將各層的分割結果進行融閤,從而通過多信息互補的方法得到完整的分割結果。設計瞭兩級集成SVM分類器對紅白細胞進行識彆。實驗證明,提齣的整套算法簡潔高效,精度高,具有較彊的普適性。
침대뇨침사도상중적홍백세포,제출료일충기우조합사상적분할방법,즉대도상진행삼층처리,장각층적분할결과진행융합,종이통과다신식호보적방법득도완정적분할결과。설계료량급집성SVM분류기대홍백세포진행식별。실험증명,제출적정투산법간길고효,정도고,구유교강적보괄성。
A segmentation method based on combination thoughts for the red and white cells in the urine sediment images is proposed,in which images are splitted into three layers,and the segmentation results of each layer are fused to get a com-plete segmentation result through multi-information complementary. Two-level integrated SVM classifier was designed to identify the red and white cells. The experiment results show that the algorithm proposed is not only simple and efficient,but also has high precision and strong universality.