利用者:Tianwei37/Report & Blog/Week 11

What I did this week:

I have achieved some substantial progress with the multi-view reconstruction UI, of course with the help of my mentor Sergey :). Basically I have separated some region-related fields (zoom, panning) in data structure SpaceClip to RegionSpaceClip, so that we can have a ‘correspondence' mode in SpaceClip. In this mode, the screen is splitted into two for users to specify correspondences across clips. The current status for this mode is that we can see a splitted view, but the video content is not correct for the witness clip.

Finish everything I should do to make a proper multi-view UI. This includes showing the right content, get the right switch, doing the version code.

There is bugs in multi-view functioning code, mainly with showing the reconstruction result of witness cameras. Sean reported this but I haven’t had time to look into it. I will dive into that section if time permits.

Based on the current status, it is highly likely I cannot deliver a mergeable branch before GSoC finishes, mainly because of the major UI changes and lacks of user testing. But I’d really like my work to be merged into master, so I’d stick with this multi-view reconstruction part to meet every standard of Blender ;).