Squashed commit of the sb-vbs branch.

Includes the SD-VBS benchmarks modified to:
- Use libextra to loop as realtime jobs
- Preallocate memory before starting their main computation
- Accept input via stdin instead of via argc

Does not include the SD-VBS matlab code.

Fixes libextra execution in LITMUS^RT.

1 files changed, 26 insertions, 0 deletions

diff --git a/SD-VBS/common/toolbox/toolbox_basic/affine/iter_AD.m b/SD-VBS/common/toolbox/toolbox_basic/affine/iter_AD.m
new file mode 100755
index 0000000..50bdae1
--- /dev/null
+++ b/SD-VBS/common/toolbox/toolbox_basic/affine/iter_AD.m
@@ -0,0 +1,26 @@
+function [A,D] = iter_AD(I,J,mask,w,k,num_trans)
+%
+%  function [A,D] = iter_AD(I,J,mask,w,k,num_trans)
+%
+%  find the affine motion A, and displacement D,
+%  such that difference between I(Ax-D) and J(x) is minimized.
+%  If k <= num_trans, only translation is computed.  This is useful
+%  in practice, when translation is relative large.
+%
+%  mask: the weight matrix,
+%  w: window size for estimating gradiant, use a large value
+%     when A,D are large.
+%
+
+%
+%  Jianbo Shi
+%
+
+
+if k <= num_trans,
+  D = find_D(I,J,mask,w);
+  A = eye(2);
+else
+  [A,D] = find_AD(I,J,mask,w);
+end
+