The task will deal with the area of Feature Extraction, in both the Frequency and Spatial domains
Read any image and identify spectral features for both radius and direction. I have applied FFT on the image and changed the coordinates to polar coordinates and used the radius and theta values as spectral features. To achieve this, run
Task2_part1.m
Extending task 1, I calculated first order histogram features, features from the co-occurance matrix from this paper. and five features from Gray Level Run Length(GLRL) matrix. I presented these features through various changes in bit-depth and direction.
To run the code, make sure you have these files in the same directory as your code:
chip_histogram_features.m
GLCL_Features1.m
glrlm.m
These are the built-in functions contributed by developers and are available on the functions repository of MathWorks.
Once you have these files in the same folder, run
Task2_part2.m
It contains the code for all three features to be run on an image.
Here is measure of all GLRL features found using glrlm.m file.