Difference between revisions of "User:Level zero/proposal"

From BRL-CAD
(Addition of working schedule)
(Edited various parts)
Line 16: Line 16:
 
|}
 
|}
  
 +
==Brief Background===
 +
I am an Undergraduate Student in Computer and Communication Engineering. Currently in my pre-final year, I have done diverse range of projects varying from Computer Systems, graphics, Image Processing and Learning.
  
===Link to Resumé===
 
 
===Brief Background===
 
I am an Undergraduate Student in Computer and Communication Engineering. Currently in my pre-final year, I have done diverse range of projects varying from Computer Systems, graphics, Image Processing and Learning. (details are in my Resume)
 
 
 
 
=Project Information=
 
=Project Information=
 
==Project Title==
 
==Project Title==
Consolidating and Adding the Image Processing tools to LIBICV
+
Consolidating and Adding the Image Processing Functions to LIBICV
 
==Brief summary of Project==
 
==Brief summary of Project==
BRL-CAD has a number of image processing tools. Currently all the tools are implemented in a modular fashion where in each tool is accessed as a module. The primitive task of this project is to combine the image processing tools to a library and advanced element of this project is to add additional relevant functions from CxImage library to the BRL-CAD’s image processing library (LIBICV).  
+
BRL-CAD has a number of image processing tools. Currently all the tools are implemented in a modular fashion where in each tool is accessed as a module. The primitive task of this project is to combine the image processing functionalities of these tools to a library. This task will ensure the reusability of these functionalities by the application programmer through relevant api calls in the code and thus will be useful for building a proper GUI for BRL-CAD and helpful in other new tools/functions to be added to BRL-CAD in future.  
This task will ensure the reusability of these tools by the application programmer through relevant api calls in the code and thus will be useful for building a proper GUI for BRL-CAD and helpful in other new tools/functions to be added to BRL-CAD in future.  
 
 
 
I have identified the following list of the various tools (current '''IPTs''') which can be integrated with the CxImage
 
 
 
 
==Project Description (Detailed)==
 
==Project Description (Detailed)==
 
===Introduction===
 
===Introduction===
*This project aims at consolidating the Image Processing Tool (IPT) (around 100 in number) under a library (already initiated as LIBICV in BRL-CAD).  
+
*This project aims at consolidating the functionalities of the Image Processing Tool (IPT) (around 100 in number) under a library (already initiated as LIBICV in BRL-CAD).  
*Additionally this project will integrate features from the CxImage library to the current tools to enhance there usability.
+
*Additionally this project will emphasize its focus on the usability of the image processing functions.
  
 
===BRL-CAD’s IPT(s) Information and Current Status===
 
===BRL-CAD’s IPT(s) Information and Current Status===
BRL-CAD’s IPT(s) are extensively written for various Image Processing task. Apart from the other formats, it has support for the BRL-CAD pioneered '''.pix''' format and the unix plots. Thus BRL-CAD cannot use any other third party library for Image Processing without altering them. Therefore work is underway in strengthening and Consolidating BRL-CAD’s image Processing Library.
+
Currently these tools are implemented as standalone applications where in each tool is a separate program and separate executable are compiled through these programs. To run any tool, currently we pass command line arguments to the executable file (separate for each tool) and the return arguments are received on standard output buffers or it could be done by saving the output at
 
+
BRL-CAD’s IPT(s) are extensively written for various Image Processing task. Apart from the other formats, it has support for the BRL-CAD pioneered '''.pix''' format and the unix plots. Thus BRL-CAD cannot use any other third party library for Image Processing without altering them. Therefore work is underway in strengthening and Consolidating BRL-CAD’s image Processing Library and adding various Image Processing functionalities in it.  
Currently these tools are implemented as standalone applications where in each tool is a separate program and separate executable are compiled through these programs. To run any tool, currently we pass command line arguments to the executable file (separate for each tool) and the return arguments are received on standard output buffers or it could be done by saving the output at some location.
 
  
 
===Consolidating IPT(s) under LIBICV===
 
===Consolidating IPT(s) under LIBICV===
This is the primary task of this project wherein the current tools are to be arranged and consolidated to a library lib_icv. It is possible to create our own library of function by writing the current function as subroutines. Any api caller just needs to include the said library and thus has access to the current modules which will be declared as functions in that library.  
+
This is the primary task of this project wherein the image processing functionalities of the current tools are to be arranged and consolidated to a  
 +
library lib_icv. It is possible to create our own library of function by writing the functions as subroutines and compiling them to link to a shared library.  
  
Libraries consist of a set of related functions to perform a common task.  In our case we have set of tools (in src/util) which are compiled separately. We can write these tools in modules (each containing one or more current tools) and compile (into objects) them individually to link that to a shared library. This could be done very easily using the make in unix which helps resolving dependencies and compile modules and link them to shared library.  
+
Any api caller just needs to include the said library and thus has access to the implemented functions of image processing.
  
Among the current tools groups will be identified which could enhance the usability of the tools. Proper Documentation will be done as discussed in later sections.  
+
Among the current tools, image functions and there groups will be identified which could enhance the usability of the tools.  
 +
Proper Documentation will be done as discussed in later sections.  
  
===Addition of library function from CxImage===
+
===Formats of Image to be supported===
====Introduction about CxImage====
+
The current status of icv.h defines .PIX,.BW,.ALIAS,.BMP,.CI,.ORLE,.PNG,.PPM,.PS,.RLE,.SPM,.SUN,.YUV and looking at the conversion tools among the IPTs in  
[http://www.xdp.it/cximage.htm CxImage] is a C++ image processing library that can load, save, display, transform images in a very simple and fast way. CxImage is open source and licensed under the zlib license.  
+
src/util folder these seem to be the only relavant formats for BRLCAD. Very few formats among them are currently implemneted for loading and saving of the
 +
images. This project will create functions for loading and saving of all these formats.
  
This library implements various image processing algorithms which are not the part of the current utilities. As per the current digging of the source code of the utilities, I propose to integrate functions from this library to the current BRL-CAD IPTs (as per the attached index) This will give added functionalities with more parameter.  
+
===Categories of Image Processing Functions===
 +
On analysing the current status of the IPTs they involve usage of the following categories of image processing functionalities.
 +
*Crop or Rect
 +
*Differeces and other airthmetic operations
 +
*Filters,
 +
*Histograms and Equalizations
 +
*Rotate
 +
*Scale or Shrink
 +
*Stats
 +
*Thresh
 +
*Interpolate
 +
*Dimensions and Image Information
 +
*Background
 +
*Borders
 +
*Fading
 +
*Merge or Split
 +
*Saturation
 +
*Morphing
  
For ''eg.'' the '''pixinterp2x''' reads and interpolates a .pix file tp double its resolution. After the project this will be integrated to the '''interpolation module''' of the '''CxImage library'''. This will make sure that the user can interpolate to arbitrary resolution using various methods thus it will enhance the usability of this interpolation function for BrlCad.  
+
In this project, I propose to add these functionalities in the icv library. Further analysis will be done to identify other functionality. During the intial
 +
working stage details of these functions will be worked with to identify the input and output arguments, usability and different methods to be incorporated
 +
in the functions.
  
====Using the CxImage Library Function====
+
===Use of image containers and structures===
 +
The current structure for used in the icv library, for image details is as follows:
 +
struct icv_image_file {
 +
    uint32_t magic;
 +
    char *filename;
 +
    int fd;
 +
    int format; /* ICV_IMAGE_* */
 +
    int width, height, depth; /* pixel, pixel, byte */
 +
    unsigned char *data;
 +
    unsigned long flags;
 +
};
 +
Like any other image processing library there will be a need of containers to keep the points, arrays etc. No, judgement is made at this stage about these
 +
structures. In the working schedule special time is devoted for this purpose. On discussion with the mentors and other developers dicision will be taken
 +
about deciding the image container and other structures needed for this library.
  
To make this library useful with the current libICV, I propose to write a bridge functions to convert between the two formats (icv_image_file struct ⇔  CxImage class). These functions may look like the following
+
===Documentation===
icv_image_file* cvt_CxImage2icv(CxImage )
+
Complete documentation will be built for all the functions. This will include the input, output arguments. Details of the various methods in the functions.
CxImage* cvt_icv2CxImage(icv_image_file* )
 
  
 
==Working Schedule==
 
==Working Schedule==
Line 107: Line 132:
 
====Sep 16 - Sep 27====  
 
====Sep 16 - Sep 27====  
 
*Final code submission with Google.
 
*Final code submission with Google.
 
==Documentation==
 
*Currently brlcad has man pages for all the tools in the utilities. I propose to change these man pages with the manuals of the functions implemented. This can be done with the sync of converting each function to the library function.
 
*Besides this since this will turn out to be a library which could be used as a third party library by users belonging to other fraternity therefore, I propose to create a complete documentation of all the functions parameters they take
 
  
 
===Why BRL-CAD?===
 
===Why BRL-CAD?===

Revision as of 22:35, 1 May 2013

Personnel Information

Name Mohit Daga
Email Address mohit.daga@ieee.org
IRC(nick) zero_level
Phone Number +91 9783582684
Mailing Address E#221 Ram Path, Shyam Nagar, Jaipur – 302019.(IN)
Time Zone UTC +0530

Brief Background=

I am an Undergraduate Student in Computer and Communication Engineering. Currently in my pre-final year, I have done diverse range of projects varying from Computer Systems, graphics, Image Processing and Learning.

Project Information

Project Title

Consolidating and Adding the Image Processing Functions to LIBICV

Brief summary of Project

BRL-CAD has a number of image processing tools. Currently all the tools are implemented in a modular fashion where in each tool is accessed as a module. The primitive task of this project is to combine the image processing functionalities of these tools to a library. This task will ensure the reusability of these functionalities by the application programmer through relevant api calls in the code and thus will be useful for building a proper GUI for BRL-CAD and helpful in other new tools/functions to be added to BRL-CAD in future.

Project Description (Detailed)

Introduction

  • This project aims at consolidating the functionalities of the Image Processing Tool (IPT) (around 100 in number) under a library (already initiated as LIBICV in BRL-CAD).
  • Additionally this project will emphasize its focus on the usability of the image processing functions.

BRL-CAD’s IPT(s) Information and Current Status

Currently these tools are implemented as standalone applications where in each tool is a separate program and separate executable are compiled through these programs. To run any tool, currently we pass command line arguments to the executable file (separate for each tool) and the return arguments are received on standard output buffers or it could be done by saving the output at BRL-CAD’s IPT(s) are extensively written for various Image Processing task. Apart from the other formats, it has support for the BRL-CAD pioneered .pix format and the unix plots. Thus BRL-CAD cannot use any other third party library for Image Processing without altering them. Therefore work is underway in strengthening and Consolidating BRL-CAD’s image Processing Library and adding various Image Processing functionalities in it.

Consolidating IPT(s) under LIBICV

This is the primary task of this project wherein the image processing functionalities of the current tools are to be arranged and consolidated to a library lib_icv. It is possible to create our own library of function by writing the functions as subroutines and compiling them to link to a shared library.

Any api caller just needs to include the said library and thus has access to the implemented functions of image processing.

Among the current tools, image functions and there groups will be identified which could enhance the usability of the tools. Proper Documentation will be done as discussed in later sections.

Formats of Image to be supported

The current status of icv.h defines .PIX,.BW,.ALIAS,.BMP,.CI,.ORLE,.PNG,.PPM,.PS,.RLE,.SPM,.SUN,.YUV and looking at the conversion tools among the IPTs in src/util folder these seem to be the only relavant formats for BRLCAD. Very few formats among them are currently implemneted for loading and saving of the images. This project will create functions for loading and saving of all these formats.

Categories of Image Processing Functions

On analysing the current status of the IPTs they involve usage of the following categories of image processing functionalities.

  • Crop or Rect
  • Differeces and other airthmetic operations
  • Filters,
  • Histograms and Equalizations
  • Rotate
  • Scale or Shrink
  • Stats
  • Thresh
  • Interpolate
  • Dimensions and Image Information
  • Background
  • Borders
  • Fading
  • Merge or Split
  • Saturation
  • Morphing

In this project, I propose to add these functionalities in the icv library. Further analysis will be done to identify other functionality. During the intial working stage details of these functions will be worked with to identify the input and output arguments, usability and different methods to be incorporated in the functions.

Use of image containers and structures

The current structure for used in the icv library, for image details is as follows: struct icv_image_file {

   uint32_t magic;
   char *filename;
   int fd;
   int format;			/* ICV_IMAGE_* */
   int width, height, depth;	/* pixel, pixel, byte */
   unsigned char *data;
   unsigned long flags;

}; Like any other image processing library there will be a need of containers to keep the points, arrays etc. No, judgement is made at this stage about these structures. In the working schedule special time is devoted for this purpose. On discussion with the mentors and other developers dicision will be taken about deciding the image container and other structures needed for this library.

Documentation

Complete documentation will be built for all the functions. This will include the input, output arguments. Details of the various methods in the functions.

Working Schedule

June 15 (Analysis and Design Period)

  • Exploring the code.
  • Make a short rough bio about each tool - the way they work.
  • Find the Image Processing functionalities used in them. Look at duplication in the functionalities.
(5-10 tools/day depending on varied length.
  • Preparation of rough ideas of each functionalities and the way they will be implemented with additional improvements in usability.
  • Discuss with mentor and other developers about the final plan for the design of the library. This includes :
    • the possible formats to be supported (In addition to the already started in include/icv.h)
    • Addition of Different functionalities and there usability improvements.
    • Discussion about common groups identified for different functionalities.
  • Learning usage from BRLCAD libraries like common.h libbu libbio

At the end of this period : A finalized plan to proceed for the development of the library.

June 17 - Aug 19 (Development Phase)

June 17 - June 24

  • Implementation of various Structures of the new library
  • Deciding about the functions to be included in different files based on categories.
  • Deciding the input parameters and return type of each function.
  • Identify error handling in each Function.
  • Witting Documentation of all the structures included. PDF+Wiki
  • Creating Patches all together.

June 24 - Aug 16

  • Implementation of image processing functionalities.
  • Witting Documentation for each functionalities. PDF+Wiki
  • Perform Unit Level Testing of all the implemented functionalities as soon as they are implemented.

July 29 - Aug 2

  • Mid Term Report

At the end of this period : A complete library with all functionalities and all of them having been tested at unit level.

Aug 19 - Sept 27 (Testing Cleaning and WrapUp)

Aug 19 - Sept 2

  • Code Cleaning
  • System Level testing for the library at multiple platforms.
  • Giving final structure to the documentation.

Sep 2 - Sep 9

  • Creating Demos For each functionality and Final cleaning

Sep9 - Sep16

  • Reserve Week.
(If everything goes right some performance analysis using gprof for different data.)

Sep 16 - Sep 27

  • Final code submission with Google.

Why BRL-CAD?

In my final year at the university, I am due to work on my Under Graduate thesis. As per my background, my university has assigned me a faculty dealing in graphics and Image Processing. This motivates me to utilise my summer holidays working with BRLCAD due to their large codebase dealing with Image Processing and Graphics and thus becoming a regular contributor with it in my final year which would help me to complete my thesis.

Why you?

I am good in c, c++. I will not say that I possess excellent coding skills. But I have enough to work on this project. Besides this I am dedicated and have an inquisitiveness to learn new things. Anything else?