Editing User:NyahCh3ck20/Proposal

From BRL-CAD

Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.

The edit can be undone. Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision Your text
Line 1: Line 1:
 
=  Personal Information =
 
=  Personal Information =
 
 
NAME : Check Nyah Watad Wallah
 
NAME : Check Nyah Watad Wallah
 
 
EMAIL: check.nyah@gmail.com
 
EMAIL: check.nyah@gmail.com
 
 
IRC:  Ch3ck
 
IRC:  Ch3ck
  
Line 10: Line 7:
 
I am a freshman Computer Engineering Student at the University of Buea, South West Region,Cameroon.
 
I am a freshman Computer Engineering Student at the University of Buea, South West Region,Cameroon.
 
After wining the 1st prize for Programming among the freshman year students at my university. I have founded a programmers club to help develop the coding skills among members of my university community
 
After wining the 1st prize for Programming among the freshman year students at my university. I have founded a programmers club to help develop the coding skills among members of my university community
 
 
== Programming Background ==
 
Languages: C(Very Good), Java(learning/Intermediate), C++(learning/basic)
 
 
C:
 
 
Wrote over a thousand lines of c code to increase performance of a sorted online dictionary used by French speaking students in our University using a redblack tree.
 
 
Implemented a fibonnacci heap to demonstrate how it can increase performance of the Dijkstra's algorithm used in solving the single-source shortest path problem used in graph search.
 
 
​Java:
 
Currently working on a Medical app to be integrated into the university website to determine the body mass Index of students in relation to the Health Awareness Initiative of the University.
 
  
 
=  Project Information =
 
=  Project Information =
Line 33: Line 17:
 
=      Detailed Project Description  =
 
=      Detailed Project Description  =
  
===        Introduction  ===
+
===        Introduction  ====
The Pull/unpush command/routine is a high priority project for BRL-CAD. The pull command seeks to restore the original state of the csg tree from any particular node after the push command has been executed. However, the Push  command is used to walk the geometry tree from a specified head node to the leaf nodes(primitive level), collecting the matrix transformations such as (translations, rotations or scales) applied to new assemblies using matrix edits(oed command). The push then applies the matrix transformation parameters to the primitives, eliminationg the need for storing the various matrix transformations thereby setting them to identity matrices. This process however looses any local coorrdinate system used in constructing the geometric objects. The pull routine seeks to restore the original tree state by reversing any tranformation operations performed on the primitive shapes from a designated top node on the csg tree. Here, I would like to show my detailed proposal in solving this summer's project. My code patch will have a sample routine that takes as argument a designated node(such as a primitive) and performs the Inverse of any rotation, Inverse of any translation and the inverse of a scale.) and performs the matrix inverse the primitive matrix.
+
The Pull/unpush command/routine is a high priority project for BRL-CAD. The pull command seeks to restore the original state of the csg tree from any particular node after the push command has been executed. However, the Push  command is used to walk the geometry tree from a specified top to the primitive level, collecting the matrix transformations such as (translations, rotations or scales) applied to new assemblies using matrix edits(oed command). The push then applies the matrix tranformation parameters to the primitives, eliminationg the need for storing the various matrix transformations thereby setting them to identity matrices. This process however looses any local coorrdinate system used in constructing the geometric objects. The pull routine seeks to restore the original tree state by reversing any tranformation operations performed on the primitive shapes from a designated top node on the csg tree. Here, I would like to show my detailed proposal in solving this summer's project. My code patch will have a sample routine that takes as argument a designated node(such as a primitive) and performs the Inverse of any rotation, Inverse of any translation and the inverse of a scale.) and performs the matrix inverse the primitive matrix.
  
  
Line 74: Line 58:
 
      0 0 1 [11]
 
      0 0 1 [11]
 
      0 0 0  1     
 
      0 0 0  1     
However, the various geometric transformations will be separated into different column matrices by the det_matrix() routine explained below.
 
  
 
The Inverse of the various transformations will be computed by the routine
 
The Inverse of the various transformations will be computed by the routine
Line 99: Line 82:
 
Inverse Translation:
 
Inverse Translation:
 
A translation is an affine transformation with no fixed points. Matrix multiplications always have the origin as a fixed point. Nevertheless, there is a common workaround using homogeneous coordinates to represent a translation of a vector space with matrix multiplication: Write the 3-dimensional vector w = (wx, wy, wz) using 4 homogeneous coordinates as w = (wx, wy, wz, 1).
 
A translation is an affine transformation with no fixed points. Matrix multiplications always have the origin as a fixed point. Nevertheless, there is a common workaround using homogeneous coordinates to represent a translation of a vector space with matrix multiplication: Write the 3-dimensional vector w = (wx, wy, wz) using 4 homogeneous coordinates as w = (wx, wy, wz, 1).
so the Inverse of a translation simply reverses the direction of the vector or matrix
+
so the Inverse of a translation is simply the reverse of the matrix. if TrnA = (a, b ,c , d)  Inverse(TrnA) = (-a, -b, -c, -d)
if TrnA = (a, b ,c , d)  Inverse(TrnA) = (-a, -b, -c, -d)
 
 
 
 
 
==The Inverse of the Transformation==
 
 
 
The key point mathematically is the reversibility of the transformations.
 
 
 
For 3D, the typical transformation is
 
 
 
[ new_x ]    [ Xx Xy Xz Tx ]  [ x ]
 
[ new_y ]  =  [ Yx Yy Yz Ty ]  [ y ]
 
[ new_z ]    [ Zx Zy Zz Tz ]  [ z ]
 
[  1  ]    [  0  0  0  1 ]  [ 1 ]
 
 
 
which is usually written as
 
v = M p
 
where v is the transformed vector, p the original vector, and M the transformation matrix (that includes rotation, scaling, and moving).
 
 
 
Because the matrix is formed as a 4-by-4 matrix as above, transformations can be combined via matrix multiplication (new matrix on the left side):
 
Mcombined = Mlater Mearlier
 
.
 
 
 
Because each transformation is reversible, there is also an inverse matrix
 
  M` M = I.
 
 
 
Thus, starting at a leaf in the object tree, the current transformation matrix on the left is multiplied with the new transformation matrix, to obtain the one that applies to the current object). The transformations are cumulative from the root to the leaf.
 
 
 
Reversing the matrix M is very simple (Xx' means transformed Xx component of the matrix, and so on):
 
D  = Determinant(M) = Zz Xx Yy - Zz Yx Xy - Zx Xz Yy - Zy Xx Yz + Zy Yx Xz+  Zx Xy Yz
 
 
 
Xx' = (Yy Zz - Zy Yz) / D
 
Xy' = (Zy Xz - Xy Zz) / D
 
Xz' = (Xy Yz - Yy Xz) / D
 
Tx' = ( -Xy Yz Tz + Xy Zz Ty + Xz Tz Yy - Xz Zy Ty - Tx Yy Zz + Tx Zy Yz) / D
 
Yx' = (-Yx Zz + Zx Yz) / D
 
Yy' = (Zz Xx - Zx Xz) / D
 
Yz' = (-Yz Xx + Yx Xz) / D
 
Ty' = (Xx Yz Tz - Xx Zz Ty + Ty Zx Xz - Yz Zx Tx - Yx Xz Tz + Yx Tx Zz) / D
 
Zx' = (Yx Zy - Zx Yy) / D
 
Zy' = (-Zy Xx + Zx Xy) / D
 
Zz' = (Yy Xx - Yx Xy) / D
 
Tz' = (-Tz Xx Yy + Tz Yx Xy + Zx Tx Yy + Zy Xx Ty - Zy Yx Tx - Zx Xy Ty) / D
 
 
 
So When the cummulative matrix is stored for each object,and its inverse any matrix transformation can be applied  in the object coordinates. This is simply done by multiplying the transformation with the object coordinates for any object; which then gives the transformation in the root coordinate system. When this is then applied to the left with each of the transformation matrices, the original transformation matrix is obtained in local object coordinates / transformations.
 
  
 
===  Overall Structure Pull routine  ===
 
===  Overall Structure Pull routine  ===
 +
The pull routine will use the inverse transformations to restore the primitive back to its original state, taking note of the local geometric transformations after each inverse transformation which will be stored together with the original matrix transformations as it moves back up the tree.
  
The pull routine works down a tree from the node pushing the directory pointers to a linked list and when it reaches the leaf, it copies the transformation matrix performs an inverse transformation on the leaf restoring it back to its initial form and moves back up the tree restoring the original tranformation matrices using the directory pointers which also stores the local coordinates.
+
However,brlcad already has the functionality for performing matrix operations on objects as in(src/libged/push.c, xpush.c). I can use the InverseTransf() as a subroutine in the pull operation and simply use the MAT_COPY directive to copy the various 4x4 original matrix transformations to the corresponding node up the tree from the leaf.
 +
So, here is a summary of the pull command
  
However,brlcad already has the functionality for performing matrix operations on objects as in(src/libged/push.c, xpush.c). I will use the InverseTransf() as a subroutine in the pull operation and simply use the MAT_COPY directive to copy the various 4x4 original matrix transformations to the corresponding node up the tree from the leaf. after the various matrix transformations have been copied from the matrix stored at the leaf node.
+
*. create a structure to hold the original matrix and corresponding local coordinate systems while calling InverseTranf().
  
So, here is a summary of the pull command
+
*. create a loop to move down the tree to the leaf nodes while recording the transformations
 
+
*. pull leaf() routine : which runs in parallel restore the original state of the primitive objects.(leafs)
*. pull_obj -create a structure to hold the original matrix and corresponding local coordinate systems while calling inverse_tranf().
 
*. det_trans()create a routine to copy the different matrix transformations performed on the primitive; which will then be given to InverseTransf() as an argument.
 
*. create a loop to move down the tree to the leaf and then recursively move up restoring the original matrix transformations.
 
*. pull leaf() routine : which runs in parallel restoring the original primitive state and copies the matrix transformation
 
which then calls the det_transf() routine to seperate the different matrix tranformations.
 
 
*. mat_restore() routine: which restores the original matrix transformation and coordinate system at each node.
 
*. mat_restore() routine: which restores the original matrix transformation and coordinate system at each node.
 +
Tests and Verification:
  
 
+
I believe test driven development would be key in finding and fixing problems with the pull routine and then with more tests to ensure the code functions correctly and its bulletproof to all forms of object inputs. This will be enhanced with the creation of a special regression test for the pull command to avoid modification of unwanted nodes/objects.Also, this command will be included among the MGED commands  and a well written manual will be made so support the usage of the command.
=== The det_trans() routine ===
 
This takes the primitives transformation matrix and separates the various types of matrix transformations from it before
 
passing the results to calling function. This routine returns an array of three 4x4 matrices having the scale, rotation and translation transformations.It returns 1 upon success and 0 otherwise.  
 
 
 
=== The inverse_transf() routine ===:
 
This routine takes any of matrix transformations determined by the det_trans() routine and determines the inverse of the  
 
transformation which will then applies to the primitive to restore its original state. It stores the results on the pull_obj structure will will then be applied to the primitive object to restore its original state.
 
  
 
=== The mat_restore() routine ===
 
This routine takes the directory arguments and corresponding transformation matrices from pull_obj structure and restores the original matrix tranformation of nodes in a particular tree.
 
 
 
=== The pull_leaf() routine ===
 
This routine applies an inverse transformation a primitive object restoring the original state or dimensions of the primitive object. returns truthfully if the object has been restored to its original state.
 
 
=== Tests and Verification ===
 
 
I believe test driven development would be key in finding and fixing problems with the pull routine and then with more tests to ensure the code functions correctly and its bulletproof to all forms of object inputs. This will be enhanced with the creation of a special regression test for the pull command to avoid modification of unwanted nodes/objects.Also, this command will be included among the MGED commands  and a well written manual will be made so support the usage of the command.
 
  
 
===  Links ===
 
===  Links ===
[1] CodePatch(Sample linked list that holds the matrices and inverses and corresponding directories together with the sample       implementation of the Inverse of a transformation function)
+
[1] CodePatch(Sample linked list that holds the matrices and inverses and corresponding directories together with the sample       implementation of the Inverse of a transformation function).
Link: https://sourceforge.net/p/brlcad/patches/169/
+
Link:  
  
 
[2] Geometric Transformations and Inverses:
 
[2] Geometric Transformations and Inverses:
https://www.cs.mtu.edu/~shene/COURSES/cs3621/.../geometry/geo-tran
+
    www.cs.mtu.edu/~shene/COURSES/cs3621/.../geometry/geo-tran.html
 +
    www.cg.info.hiroshima-cu.ac.jp/~miyazaki/knowledge/teche53.html
  
http://www.cg.info.hiroshima-cu.ac.jp/~miyazaki/knowledge/teche53
 
  
 
= Deliverables =  
 
= Deliverables =  
Implementation of the inverse_tranf(), mat_restore(), det_transf() routines and other pull subroutines.
 
 
Implementation of the complete pull routine with further testing.:(pull_leaf())
 
  
Integration  of the Pull routine into MGED command interface together with Documentation containing(summary together with usage capabilities)
+
    Implementation of the Inverse transformation routine. and other pull subroutines(mat_restore()).
 +
    Implementation of the complete pull routine with further testing.
 +
    Integration  of the Pull routine into MGED command interface together with Documentation containing(summary together with usage capabilities)
 +
   
  
 
= Development schedule =
 
= Development schedule =
===July 1st (~ 3 weeks)===
 
Study of BRLCAD Manuals and other Documentation on the Push Command
 
 
Discuss with other developers concerning implementation details.
 
 
Study the (src/libged , /include ) libraries and the implementation of push/xpush commands.
 
 
Discuss more coding specifications and implementations details with mentor/other developers
 
       
 
===July 21 - July 28 (1 week)===
 
Implementation of the det_trans() and InverseTransf() routines with
 
determination of original matrix transformations.
 
  
Tests on sample primitive matrices
+
== July 1st (~ 3 weeks) ==
       
+
        Study BRLCAD Manuals and other Documentation on the Push Command
===July 29 - Aug 17(3 weeks)===
+
        Discuss with other developers
Implementation of do_restore() routine(1 week)
+
        Study the (src/libged , /include ) libraries and the implementation
which traverses nodes restoring the original matrix and tests
+
        push/xpush commands.
 
+
        Discuss more coding specifications and implementations details with 
Implementation of pull routine(pull_leaf() and others)(2 weeks)
+
        mentor/other developers
 
+
==July 21(1 week)
Testing and functionality verification of function.
+
        Implementation of the Inverse Transformations with
 +
            determination of original matrix transformations
 +
        Tests on sample primitive matrices
 +
==July 28 - Aug 18(4 weeks) ==
 +
Implementation of do_restore() routine
 +
which traverses nodes restoring the original matrix
 +
        Implementation of pull routine(pull_leaf() and others)
 +
          Testing and functionality verification of function
 
            
 
            
Mid-term evaluation in July 29 - Aug. 2  
+
        Mid-term evaluation in July 29 - Aug. 2  
 
          
 
          
===Aug. 18 - Aug. 31 (2 weeks)===
+
== Aug. 19 - Sept 14 (4 weeks) ==
Finalization of complete pull routine  
+
        Finalization of complete pull routine  
 
+
       
Tests
+
        Tests
Tests of the final pull routine on primitives  
+
            Tests on the final pull routine on primitives  
   
+
           
===Sept 01 - Sept 14( 2 weeks)===       
+
        Integration of Pull into MGED command interface.
Integration of Pull into MGED command interface.
+
            Testing of functionality of command  
 
+
            and debugging
Testing of functionality of command and debuggging
+
== Sept 15 - Sept 21(1 week) ==
   
+
    Tests
===Sept 15 - Sept 21(1 week)===
+
    Fix bugs and improve performance of routine
Tests
+
    Documentation and code clean up
Fix bugs and improve performance of routine
 
 
 
Documentation and code clean up
 
 
    
 
    
===Sept 23 - Sept 27(1 week)===
+
== Sept 23 - Sept 27(1 week) ==
Final Evaluation
+
    Final Evaluation
Submission of Final code to Google.
+
    Submission of Final code to Google.
  
 
= Time availability =
 
= Time availability =
I would be able to offer over 40 hours on the project. However, Our Second Semester ends late june or early july and our next semester begins in early October. However, if the semester extends to early july it will be for completion of exams so i would dedicate most of my evenings and weekend in the first week of July to start work on studies of the brlcad documentation and libraries.
+
 
 +
I would be able to offer over 40 hours on the project. However, Our Second Semester ends late june or early july and our next semester begins in early October.
  
 
= Why BRL-CAD? =
 
= Why BRL-CAD? =

Please note that all contributions to BRL-CAD may be edited, altered, or removed by other contributors. If you do not want your writing to be edited mercilessly, then do not submit it here.
You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource (see BRL-CAD:Copyrights for details). Do not submit copyrighted work without permission!

To edit this page, please answer the question that appears below (more info):

Cancel Editing help (opens in new window)