利用者:Apinzonf/Gsoc2013/progress

GSOC 2013 Progress Reports

Week 1(jun 3 - 9)Report

What I did this week?

I read this papers and documents:

• Laplacian surface editing (Sorkine, O.; Cohen-Or, D.; Lipman, Y.; Alexa, M.; Rossl, C. & Seidel, H.-P. Laplacian surface editing Proceedings of the 2004 EurographicsACM SIGGRAPH symposium on Geometry processing, ACM, 2004, 175-184)
• Linear Rotation-Invariant Coordinates for Meshes (Lipman, Y.; Sorkine, O.; Levin, D. & Cohen-Or, D. Linear rotation-invariant coordinates for meshes ACM Trans. Graph., ACM, 2005, 24, 479-487)
• Master Thesis Object Deformation Using Laplacian Coordinates. (Alvaro Perez)

I implement a Laplacian Editing 3D in matlab

• LaplacianEditing3D.m

What you plan to do next week?

Create experimental code for test laplacian editing in Blender source

Any roadblocks, difficulties, etc.

The deformation depends on the numerical method used

Week 2(jun 10 - jun 16) Report

What I did this week?

I spoke with Sergey and helped me add the development branch soc-2013-sketch_mesh

https://svn.blender.org/svnroot/bf-blender/branches/soc-2013-sketch_mesh/

Any roadblocks, difficulties, etc.

.

Week 3(jun 17 - jun 23) Report

What I did this week?

This week I started to program the proposal as a bmesh operator.

The code I wrote to be uploaded is unstable, but here's the patch to see the content.

[Soc-2013-sketch_mesh-21-06-2013.patch]

I wrote part of the design document

User:Apinzonf/Gsoc2013/design

What you plan to do next week?

Implement the problem with Eigen3 libraries to solve the linear system.

To finish writing the design document

Any roadblocks, difficulties, etc.

I had problems with mixing C code and C + + because the library to solve the linear system is SuperLU (C) and Eigen3 (C++) for matrix operations.

I have another big problem and until now I see SuperLU works only with NxN square matrices, and the matrix that I have is of MxN.

Week 4 (jun 24 - jun 30) Report

What I did this week?

I did my first commit, I use a long time adjusting the equation for calculating the rotations implicit. I did implement a method to rotate the differential coordinates based on paper Linear rotation-invariant coordinates for meshes (Lipman 2005)

What you plan to do next week?

I will talk with my mentor Howard Trickey and artist Jonathan Williamson, to define the user interaction of the tool. To finish writing the design document

Any roadblocks, difficulties, etc.

I had many problems with the manage of indexes, and the operations with floating point data.

Week 5 (jul 1 - jul 7) Report

What I did this week?

This week did not work a lot, was the feast of San Pedro and went on vacation to a nearby town.

I finished writing the design document.

User:Apinzonf/Gsoc2013/design

What you plan to do next week?

Write the code to mark the vertices as static and handlers, according to the interaction with the user defined.

Any roadblocks, difficulties, etc.

I have an inflammation in the left cornea (keratitis).

Week 6 (jul 8 - jul 14) Report

What I did this week?

This week I implemented the deformed Laplacian as a tool EditMesh code based on the knife tool. This tool will allow the user to interact with the mesh while modeling

What you plan to do next week?

I will improve the code, and correct any problems found.

I will discuss with my tutor about possible forms of interaction with the user

Any roadblocks, difficulties, etc.

.

Week 7 (jul 15 - jul 21) Report

What I did this week?

This week I've fixed some problems of the tool:

• Previously used two contexts SuperLU, Now only used a context of SuperLU.
• The system factorizes the Laplacian matrix only one time, so when you modify the vertices only need to calculate the solution without the need to factor again.
• I remove the computation of implicit rotations because not improve the solution.
• I eliminate the use of the library Eigen3 in my system because SuperLU library has better performance.

I built one in windows version to download and test.

• This is a test version has some bugs and console displays debug messages.
• This is a work in progress and the next version will use the grease pencil.
• This system works with meshes composed of triangles or quads. (It is not a trivial task, and academic papers solve it only in mesh with triangles).
• The user interface is not finished.
• I wrote page on the Wiki, on that page explains how to use the tool.

User:Apinzonf/Gsoc2013/testing

I publish this page in http://www.blenderartists.org/, so that people begin to test and review

http://www.blenderartists.org/forum/showthread.php?300973-GSOC-2013-Sketch-Mesh-Editing-Video-Test-and-Downloads&p=2425592#post2425592

I made a video showing what can be done now with the application at this time

What you plan to do next week?

I will analyze the comments received on the website of BlenderArtist, to propose a design for user interaction.

I will continue investigating the problems with the system when the meshes are dense.

Any roadblocks, difficulties, etc.

I have had many problems with the size of the edges and floating point, I'm still researching how to improve it.

.

Week 8 (jul 22 - jul 28) Report

What I did this week?

Considering the thread some opinions on the site user Blenderartist as Indy_logic and David Jeske(davestf), i implement the deformed Laplacian as a modifier.

In the Laplacian Deform modifier, used two vertices vertexGroups one for static and one for vertices handlers.

By implementing the tool as Modifier allowed the use of other modifiers as Hook, so that the vertices handlers can be controlled by various Hook objects.

Blenderartist thread:

http://www.blenderartists.org/forum/showthread.php?300973-GSOC-2013-Sketch-Mesh-Editing-Video-Test-and-Downloads

What you plan to do next week?

I will improve the code by the switch that has some memory management problems.

Any roadblocks, difficulties, etc.

When the meshes are dense the nonlinear optimization system fails. The academic paper used not mention these problems, for that reason the system will have some limitations.

Week 9 (jul 29 - ago 04) Report

What I did this week?

This week may not work on the project; i was busy working on the presentation of the paper “Shape with an Adapted Laplacian Inflation Operator For Hybrid Quad / Triangle Meshes”. The paper will be presented at the conference that will be held SIBGRAPI 2013 5 to August 8, 2013 at Arequipa Peru.

That article was one of the results of my work in the GSOC 2012. Mesh smoothing based on curvature flow operator in a diffusion equation. Blender Foundation.

I send a copy to those interested in the article will use blender as the test platform, and all 3d images were made with Blender.

Link to download the PDF

This is a small contribution in Computer Graphics as a result of joining the academy and the Blender Foundation

What you plan to do next week?

Next week I will travel to Arequipa Peru to present the paper at the conference SIBGRAPI.

Any roadblocks, difficulties, etc.

.

Week 11 (ago 12 - ago 18) Report

What I did this week?

The modifier now only requires a group of vertices that will be the handlers. The system no longer requires the use of static vertices.

This week also started to set Bmesh operator, so you can capture the silhouette of the vertices to modify based on a simple strategy in object space.

What you plan to do next week?

Next week I will to mark handlers with base in silhouette detected in object space.

Any roadblocks, difficulties, etc.

.

Week 12 (ago 19 - ago 25) Report

What I did this week?

I implement a simple method for capturing the silhouette based on an algorithm that works in the object space. In the implemented method the user selects a region of interest. In this region the silhouette is detected, and the vertices in this silhouette will be the handlers. The vertices that are not in the region of interest are marked as static vertices.

What you plan to do next week?

I goint to add the algorithm to work with orthographic view

Any roadblocks, difficulties, etc.

I have problems with event handling in blender when I select some vertices from the code in blender and the system to try to rotate or scale it falls.

.

Week 13 (ago 26 - sep 1) Report

What I did this week?

This week i can’t advance in my project because the National University of Colombia had been block by manifest from syndicate. In this university I have a computer for work in gsoc project. Also the city where I live Bogota has had security problems due to violent disturbances due to agricultural strike in Colombia.

Week (nov 18 - nov 24) End Report

I created a patch for master.

This patch only include Modifier, and code reviewed

I put patch in Differential in

http://developer.blender.org/D34