Cyclops Release #10
Well, its been over six months since the last update. First there was the thing alluded to in the last post, then I broke my arm (Badly, like multiple fractures and a whole load of metal badly.) and now I'm in the last year of my PhD, which makes time something of a commodity. This weekend has proved free however and I thought I'd spend some of it polishing Cyclops just enough to do another release.
Despite the above there are a few new features in this release however, and some are actually useful:
* The Mesh Wibbalizer tool has been added. This allows you to load and save 3D models and apply various basic operations to them. The most useful is the Simplify tool which makes the meshes saved by the 'Disparity To Mesh' tool far easier to use. The Factor/Vertices entry box can either be set to less than 1 to get that percentage of the vertices in the original mesh, or larger than 1 to simplify down to a precise number of vertices. Note that its not perfect, I'm sure there are bugs in there, but I got tired of hunting them as they do not manifest when its used on meshes output by the tool, only meshes created elsewhere. (By bugs I mean it very occasionally creates degenerate geometry. But not for the meshes Cyclops creates, so it does what I want.)
* Triangulator has had a numerical error bug fix that could shift output coordinates slightly from where they should be. Its also gained the ability to reload the 3D models it saves out and re-project them, so you can save and come back to working on one latter. It also allows you to edit in a 3D modelling tool and then reload - re-projection recalculates all the UV coordinates, compensating for any distortion you get from moving vertices around in the modelling tool.
* There are three new sub-tools all dedicated to shape from shading. (Calculating the shape of an object using its shading.) These came about due to a paper I just submitted, and include my algorithm plus two others - Lee & Kuo and Worthington & Hancock. I'm not going to reveal anything about my algorithm until its actually published, which seeing as I've submitted to a conference with a 1 in 5 acceptance rate is probably not going to be any time soon.
* Fixed a massive bug in the intrinsic calibration tool that meant it crashed. The fact that has been in there for 6 months and nobody noticed gives me all the information I need to know about how many users I have...
* Manual has had some improvements, though its still out of sync with the program. I\'ll hopefully get some more work in over the next month, at least enough to make sure its current contents matches the actual program.
Well, to my zero users: Enjoy!
Despite the above there are a few new features in this release however, and some are actually useful:
* The Mesh Wibbalizer tool has been added. This allows you to load and save 3D models and apply various basic operations to them. The most useful is the Simplify tool which makes the meshes saved by the 'Disparity To Mesh' tool far easier to use. The Factor/Vertices entry box can either be set to less than 1 to get that percentage of the vertices in the original mesh, or larger than 1 to simplify down to a precise number of vertices. Note that its not perfect, I'm sure there are bugs in there, but I got tired of hunting them as they do not manifest when its used on meshes output by the tool, only meshes created elsewhere. (By bugs I mean it very occasionally creates degenerate geometry. But not for the meshes Cyclops creates, so it does what I want.)
* Triangulator has had a numerical error bug fix that could shift output coordinates slightly from where they should be. Its also gained the ability to reload the 3D models it saves out and re-project them, so you can save and come back to working on one latter. It also allows you to edit in a 3D modelling tool and then reload - re-projection recalculates all the UV coordinates, compensating for any distortion you get from moving vertices around in the modelling tool.
* There are three new sub-tools all dedicated to shape from shading. (Calculating the shape of an object using its shading.) These came about due to a paper I just submitted, and include my algorithm plus two others - Lee & Kuo and Worthington & Hancock. I'm not going to reveal anything about my algorithm until its actually published, which seeing as I've submitted to a conference with a 1 in 5 acceptance rate is probably not going to be any time soon.
* Fixed a massive bug in the intrinsic calibration tool that meant it crashed. The fact that has been in there for 6 months and nobody noticed gives me all the information I need to know about how many users I have...
* Manual has had some improvements, though its still out of sync with the program. I\'ll hopefully get some more work in over the next month, at least enough to make sure its current contents matches the actual program.
Well, to my zero users: Enjoy!