Ask Your Question

hog+svm: the size of positive image must be 64*128? any other choice?

asked 2013-12-27 01:53:04 -0600

kelvin00 gravatar image

updated 2013-12-27 01:54:22 -0600

Iam trying to train my own SVM using features extracted using HoG. Now i have one question. should the size of positve image must be 64x128? if the positive pictures are small ,and i can resize them to 64*128. however,the targets are can i build the positive picture? Dose the positive picture only include the target? thanks a lot. image description

edit retag flag offensive close merge delete

1 answer

Sort by » oldest newest most voted

answered 2013-12-27 04:17:26 -0600

The sample size of a classifier has to be carefully chosen accordingly to the problem. For instance, the 64x128 size is commonly used in pedestrian detection, since a pedestrian image has obviously more height then width. Having that said, 64x128 isn't a good size for detecting your object as its aspect ration is wider. You must create a training set with your chosen sample size, to then perform detection with that same size.

You could even try 128x64 at first. It would be easier to implement because it is very similar to the widely used 64x128, and then if performance is not good enough, try to tweak it until you are satisfied

edit flag offensive delete link more


thank you very much, Median. I'll have a try to use the 128X64 and tell you the result.

kelvin00 gravatar imagekelvin00 ( 2013-12-29 19:06:07 -0600 )edit

Question Tools


Asked: 2013-12-27 01:53:04 -0600

Seen: 1,547 times

Last updated: Dec 27 '13