模式识别RVM_PR_yc幻灯片.pptVIP

Test and clusters identification the trained classifier model is applied to classify each pixel location as “MC present” and “MC absent”; 2) next, the detected “MC present” pixels are grouped into potential MC objects by a morphological processing procedure, during which isolated spurious pixels are removed; 3) the detected MC objects are grouped into clusters using a clustering criterion. A true positive (TP) cluster : the objects are connected with nearest-neighbor distances (Dnn) less than 0.4 cm; 2) at least three true MCs should be detected by the algorithm with an area of 1 cm2. A false positive (FP) cluster if the objects satisfy the above conditions but contain no true MCs. Result The detection performance is evaluated using the free-response receiver operating characteristic (FROC) curve Result Computational complexity Implementation was in MATLAB on a 2.13 GHz Intel PC with 3GB memory and Windows 7 operating system. Training time (sec) Testing time (sec) RVs aRVM 3,572.3 8.7 5 RVM 2,198.4 28.6 72 Conclusion The aRVM classifier can achieve better detection performance than the RVM classifier when the false-positive rate is low. It also yields a much sparser model, which has lower computational complexity, making it more desirable for real-time processing of mammogram images. 乳腺X线图像中微钙化点簇的检测 Detection of Clustered Microcalcifications in Mammograms Presentation ? 2012 BJTU BEIJING JIAOTONG UNIVERSITY 乳腺X线图像中微钙化点簇的检测 Detection of Clustered Micro-calcifications in Mammograms 姚畅 chyao@bjtu.edu.cn 北京交通大学 2012-3-17 * . * . 内容提要： 一、相关向量机(Relevance Vector Machine, RVM)理论 二、基于自适应核学习的乳腺微钙化点簇检测 * 乳腺X线图像中微钙化点簇的检测 Detection of Clustered Microcalcifications in Mammograms . * . Relevance Vector Machine (RVM)： 与support vector machine (SVM) 的关系： -- 都是通过对训练样本集进行训练，获得训练好的模型，实现对新数据的预测。 SVM： -- The number of support vectors grows linearly with the size of the training set. -- Predictions are not probabilistic. -- Necessary to estimate the error/marg