Extended study on cleaning gaussian noise in mechanical engineering drawing images using median filter and its variants
Loading...
Date
2008-06
Authors
Khong Teck, Low
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Vectorization is a process that converts a raster image into a corresponding
vector image. The quality of the detected vectors is based on how accurate the
attributes of the lines in the raster image were recognized by the vectorization
methods. The detected vectors may have some distortions due to many operations
that have been operated upon the raster images. Among the operations that may
affect the quality of the detected vector are the amount of noise, the noise cleaning
method, and the vectorization software used. In this research, we perform an
extended study on the effect of different factors on the quality of vector data based
on a previous study.
In the noise factor, we will study one kind of noise that appears in document
images namely Gaussian noise while the previous study involves only salt-andpepper
noise. High and low levels of noise are studied. For the noise cleaning
methods, we use algorithms that were not covered in the previous study namely
Median filters and its variants.
For the vectorization factor, one of the best available commercial raster to
vector software namely VPStudio is used to convert raster images into vector format.
The performance of line detection will be judged based on objective performance
evaluation method.
The output of the performance evaluation is then analysed statistically to
highlight the factors that affect vector quality.
Description
Keywords
Cleaning gaussian noise , Mechanical engineering drawing images , Median filter