Lab: Image Sharpening

[Exercises] [Workspace]

Purpose: The purpose of this lab is to investigate image sharpening using high-pass filtering techniques.

Activities:

1. Read the image $DIP/data/retina.kdf, convert its data type to "float", and display it, using operators User defined, Convert Type, and Display Image.

   1. Glyphs:Input/Output:Data Files: User defined

   2. Glyphs:Data Manip:Data Conversion: Convert Type

   3. Glyphs:Visualization:Non-Interactive Display: Display Image

2. Perform the convolution of the image and the Laplacian kernel contained in the file $DIP/data/laplacian.ascii, and display the result. Use operator LinearOp (Linear Operator) to perform the convolution.

   1. Glyphs:Input/Output:Data Files: User defined

   2. Glyphs:Arithmetic:Linear Transforms: LinearOp (Linear Operator)

   3. Glyphs:Visualization:Non-Interactive Display: Display Image

3. Pad the kernel and take its Fourier transform to visualize its response in the frequency domain. The operators are Pad and FFT. Try adding the "SA PSeudo" colormap to the spectrum prior to displaying it using the Autocolor operator.

   1. Glyphs:Data Manip:Size & Region Operators: Pad

   2. Glyphs:Arithmetic:Linear Transforms: FFT

   3. Glyphs:Visualization:Map Display & Manip: Autocolor

   4. Glyphs:Visualization:Non-Interactive Display: Display Image

4. Add the original image with its filtered version and display result.

   1. Glyphs:Arithmetic:Two Operand Arithmetic: Add

   2. Glyphs:Visualization:Non-Interactive Display: Display Image

Exercises

1. Design a combined kernel which performs the laplacian and the addition simultaneously.
2. Perform the same experiment, but modifying it to use high-pass kernels.

Khoros Workspace
Execute the visual program c6s4laplacian-sharp.wk