Reputation: 3497
What I have tried so far...
opencv_createsamples -info positive/positive.txt -vec positive.vec -w 100 -h 100 -num 2
opencv_createsamples -info negative/negative.txt -vec negative.vec -w 100 -h 100 -num 2
opencv_traincascade -vec positive.vec -data . -bg negative/negative.txt -numPos 1 -numNeg 1 -numStages 1 -w 100 -h 100
# outputs
Train dataset for temp stage can not be filled. Branch training terminated. Cascade classifier can't be trained. Check the used training parameters.
How to train a HAAR model with very few positives? The goal is to track 2d images in 2d images.
Upvotes: 4
Views: 693
Reputation: 3236
Per http://docs.opencv.org/2.4/doc/user_guide/ug_traincascade.html#negative-samples negative samples can't be created by createsamples. Unfortunately, the documentation is not super clear about why you can't, and the question hasn't been clear about what you expect this to do.
Upvotes: 1
Reputation: 3236
One of your arguments is "numpos" but per http://docs.opencv.org/2.4/doc/user_guide/ug_traincascade.html, it should be "numPos" with a capital P.
Upvotes: 1