ishit/l1stabilizer

:movie_camera: Video stabilization using L1-norm optimal camera paths.

l1stabilizer's Issues

About script content

Hello, I would like to ask what is the specific content of the script'/vlfeat/toolbox/vl_setup.m' and '/cvx/cvx_setup.m' and where are they exactly

the meaning of each element in c1

1. in optimizeTransforms.m,
   % Weighting of 100:1 for affine to translational parts
   c1 = [1 1 100 100 100 100]';
   Even though you said weighting of 100:1 for affine to translational parts, i still can not understand the meaning of each element in C1, can you explain it?

2. when i run 'main.m' for my own video, but the new camera path that i get is horizontal line in figure 1 or figure 2 about the movement of x or y axis.
   can you tell me how to adjust the parameter so that the movement of x or y axis for new camera path is match the movement of x or y axis for original camera path?

Thanks a lot

Does this work support real-time stabilization?

Does this work support real-time stabilization?
Thanks.

optimized path is approximate linear

hi, thinks for your sharing. now i met a problem. when plotting optimized path, it is approximate linear, and your project is perfect. What is the cause of this?

Incorrect use cvx/subsref

Hello, I always encounter the following error:
Incorrect use cvx/subsref (line 13)
The subscript index must be a positive integer type or a logical type.
p(k, 6) = p(n, 6);
n_transforms = optimizeTransforms(t_transforms, im_size);
I want to know how to solve.Thank you!

The issue about the Saliency Constrains

After reading the code,I find that the algorithm did't add saliency constraints according the paper, I have try to recurrent this part of the code, but I encountered some problems now, and I want to know if you have recurrented this part code? If you have finished it, can you share the code? If the code is unavailable, can you share some ideals?

thanks a lot!