opencv different results on win vs linux

asked 2016-04-25 23:07:17 -0500

I ran the same code on windows and linux. But the results are totally different.

First, I used adaboost to learn the model. On win, it runs well. On linux, when training adaboost, there is an error:

Bad argument (168-th value of 4-th (ordered) variable (=-inf) is too large) in CvDTreeTrainData::set_data, file /build/buildd/opencv-2.4.8+dfsg1/modules/ml/src/tree.cpp, line 551 terminate called after throwing an instance of 'cv::Exception' what(): /build/buildd/opencv-2.4.8+dfsg1/modules/ml/src/tree.cpp:551: error: (-5) 168-th value of 4-th (ordered) variable (=-inf) is too large in function CvDTreeTrainData::set_data

Then, I used SVM to replace adaboost. On win, the result is OK, however, the result on linux is awful. I use the same code and same paremeters.

How can I deal with this issue.

edit retag flag offensive close merge delete