todo.txt

List of To Do's:

- Fix txbr_bin.py if length is not a multiple of 2
- Fix the combining C code (remove unuseful memcp)
- Fix the backprojection C code
- Check Matlab vs C indexing condition
- pb with the rotation axis 
- pb with starting from different .txbr file
- trace angle only on the smaller tilts
- Filtering angle sign
- Stacked series and rotation axis
- Make sure the number of tilts in rawtilt is the same as the number of images
- Do a plotting for the angles in an orthogonal approximation
- In contour alignement, contours for an extended cannot be empty->validation
- Remove zzopencv dependencies
- Need to hook-up the scaling evaluation
- Fix the relax projection routine scheme in the case of the orthogonal approximation
- Problem with contour alignment and n2=2
- Problem initialization multiple series (with the sample transfer)
- Check the number of blocks pb in the cases: back.2/  back.cu/  back.cu.2/

Current
- Fix the triton txbr_finalize.py bug. On a cluster, the starting directory is always the top
directory on each node. Always use absolute path.

Add-ons and Corrections:

v3.0.4
- Add magnification variables in the orthogonal multiple beam approximation for the alignment. Symmetric case assume the same magnification factor in both direction of the camera. This is the default option.

v3.0.2
- Make the interpolation mode of the flattening work on top of a polynomial approximation
- Make the interpolation mode works on a fix size frame (X_INTERP,Y_INTERP) for performance
- Make the runtxbr.py program stops if one of the step did not work properly (20-10-2009)
- Ability to read the txbr.cfg file from the work directory (default value ".")

v3.0.1
- Make sure the reorientation of the specimen is done after "on the fly" flattening: the problem was elesewhere... Reorientation of the specimen is intrinsically done within the "on-the-fly" routine. There was a bad rawtilt files with a different numbers of angles compared to the preali file (19-10-2009).
- Orthogonal approximation is broken: projection map coefficients were shifted, unfold_var2_with_scaling was not calling unfold_var2