proposal for

Evaluation of Image Processing Filters

 

submitted to fulfill the semester project requirement for

 

EE4773/6773: Digital Signal Processing

 

September 15, 1997

 

submitted to:

 

Dr. Joseph Picone

Department of Electrical and Computer Engineering

Mississippi State University

MS39762

 

submitted by

 

Brandon Butler, Nivedita Sehasrabudhe, Phani Mantena

and

Shivaraj Tenginakai

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I. Abstract

The goal of this project is to evaluate various filters used in image processing. The filters to be evaluated include examples from both interpolation and derivative class of filters. The basic design involves taking in data from regularly sampled 3D data sets and reconstructing the image using the filters under study. The evaluation criteria is based on the standard methodologies present in the literature.

II. Introduction

Background

In volume rendering, the data to be rendered is in the from of a large array of numbers, sampled at discrete set of points in space. A function capable of providing a value for every point in the volume is necessary to perform many operations vital to visualization such as simulating light propagation or extraction of isosurfaces.

Filters are functions used to reconstruct the entire volume space, using information gathered at these set of points. The filters considered here deal with regular sampling, in which the data set consists of points sampled at regular intervals in the lattice.

The two basic classes of filters are the interpolation filters and the derivative filters. As stated earlier, filters from both the classes shall be evaluated in this project.

Test Filters

The various interpolation filters used for evaluation include (Marschner 1994) :

    1. The Trilinear filter

    2. The Catmull – Rom spline filter

    3. The Gaussian filter

    4. The Windowed 3-sinc filter

The various derivative filters used for evaluation include (Moller 1996) :

    1. The Cubic derivative filter

    2. The Gaussian gradient filter

The project design will be flexible enough to allow the addition of new filters when needed.

The Evaluation Criteria

The various filters shall be evaluated for the most important filter qualities:
    1. Smoothing
    2. Postaliasing
    3. Ringing

A plot of the various filters for each of the above evaluation criteria shall be obtained, as the final result of this project.

III. The System Design

The System diagram is shown in the following figure,


    Database ---->  FILTER MODULE    ----> EVALUATION MODULE

                          |                   |
                          |                   |

                    RENDERING MODULE ---->   GUI

As seen from the diagram there are four modules that need to be designed,

  1. Filter Module - this module simulates the various filters which are used in evaluation.

  2. Evaluation Module – this module implements various evaluation criteria.

  3. Rendering Module – this module renders the image represented in the database.

  4. GUI – provides the user interface and coordinates the various activities.

The Database is taken from external sources.

 

 

 

 

 

VI. The Schedule

The following is the schedule for the completion of this project

 

Tasks

 

Deadline

Proposal

September 15

Research & Design Freeze

September 25

Filter

Module

September 31

Evaluation and Rendering Module

October 15

GUI

October 20

Integration

October 25

Compilation of Results

October 30

Presentation and Documentation

November 5

VII. REFRENCES

Marschner, Stephen R, and Richard J. Lobb. 1994. An evaluation of reconstruction filters for volume rendering. IEEE: 100-107.

Moller, Torsten, Raghu Machiraju, Klaus Mueller, and Roni Yagel. 1996. Classification and local error estimation of interpolation and derivative filters for volume rendering. IEEE: 71-78.