-1

In the section of attentional cascade they have discussed a basic 2 feature strong classifier. While discussing the computation of this weak classifier they required

  1. Evaluate the rectangle features (requires between 6 and 9 array references per feature).
  2. Compute the weak classifier for each feature (requires one threshold operation per feature).
  3. Combine the weak classifiers (requires one multiply per feature, an addition, and finally a threshold).

My question is where do we require the multiplication. I supposed that the result of each weak classifier function h(t) is either 1 or 0. So we either multiply alpha with a 1 or a 0 which is decision rather multiplication.

Or may be I am not exactly getting the algorithm

Thanks

user2793078
  • 419
  • 7
  • 12
  • 1
    You are referencing a document you have not supplied a link for or attached. How are we supposed to know what 'In the section of attentional cascade they have discussed a basic 2 feature strong classifier.' refers to?? You will need to update the question accordingly!! – MoonKnight Nov 14 '13 at 15:48

1 Answers1

1

I guess you refer to: http://www.cs.cmu.edu/~efros/courses/LBMV07/Papers/viola-cvpr-01.pdf

Weak classifier are combined using a weighted sum of their results. The multiplication occurs when multiplying the result (0 or 1 in your case) with the weight.

See the formula on page 4 below The final strong classifier is

Simon Bergot
  • 10,378
  • 7
  • 39
  • 55
  • I am referring the paper http://www.cs.cmu.edu/~efros/courses/LBMV07/Papers/viola-IJCV-01.pdf . Here at page 12 first line "Combine the weak classifiers (requires one multiply per feature, an addition, and finally a threshold)." I just want to confirm can we simply replace the multiplication with decision block which decided whether to add the weight (multiplication with 1) or not(multiplication with 0). Just confirm my thinking – user2793078 Nov 14 '13 at 19:27