APPLICATION OF THE LASSO FUNCTION IN COMPRESSIVE SENSING FOR IMAGE RECONSTRUCTION OF MICROWAVE TOMOGRAPHY
Tomography is a technique for observing the inside of an object without having to damage the object. Microwave tomography is one of tomographic techniques using microwave frequencies with the advantage such as a relatively small antenna of used device. This study conducted a case study of the app...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/54478 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Tomography is a technique for observing the inside of an object without having to
damage the object. Microwave tomography is one of tomographic techniques
using microwave frequencies with the advantage such as a relatively small
antenna of used device. This study conducted a case study of the application of the
least absolute shrinkage and selection operator (LASSO) function in the
compressive sensing (CS) method for microwave tomography image
reconstruction applications and analyzed its performance with image quality
parameters. LASSO is a reconstruction algorithm in CS which is a linear model
with ????1 regularization. The research consisted of image processing with test
image and image processing with measurement data. Image processing with test
image using a photo of a tree trunk object with different pixel sizes. Image
processing from the measurement data uses the signal results captured by the
receiving antenna after passing through a hollow wooden tree object.
Performance was assessed using the mean square error (MSE), peak signal to
noise ratio (PSNR), and structural similarity index measure (SSIM) parameters
against the ???? value of the LASSO function in CS. In general, the image
reconstruction results show the best quality values at the smallest ????. The
reconstruction results for image processing with test image at 256 × 256 pixels
have the best MSE value of 0.001, the best PSNR value of 78.130 dB, and the best
SSIM value of 0.944 with the longest computation time of 2 hours 11 minutes 53
seconds. While the reconstruction results for signal processing with measurement
data at a frequency of 2.885 GHz in a size of 24 × 24 pixels have the best MSE
and SSIM with values of 0.1293 and 0.1838 with an average computation time of
30 seconds for one test. |
|---|