Hadamard Transform Imaging Paul Holcomb Tasha Nalywajko Melissa Walden.
-
Upload
russell-francis -
Category
Documents
-
view
216 -
download
0
Transcript of Hadamard Transform Imaging Paul Holcomb Tasha Nalywajko Melissa Walden.
Problem Definition
• Current 3D spectral imaging systems are relatively slow and possess low resolution
• Hadamard multiplexing greatly enhances signal to noise ratio, thus enhancing resolution
• Use of a digital micro-mirror device makes implementation of the Hadamard transform very fast
Hypothesis
Implementation of the Hadamard transform in a spectral imaging system by using a digital micro-mirror device will decrease the amount of time
necessary to acquire an image and increase the spatial and spectral resolution of the image.
Faster, more accurate spectral imaging will allow for better differentiation of tissue during brain
tumor resectioning and increase survival rates.
3D Spectral Imaging Techniques - Scanning
Wavelength Scanning
http://www.cri-inc.com/instruments/data/pdf/American%20Laboratory%202000.pdf
Line Scanning
http://www.lightforminc.com/Technology/ANIMATION/animation.html
3D Spectral Imaging Techniques - Multiplexing
Fourier Transform
http://scienceworld.wolfram.com/physics/FourierTransformSpectrometer.html
Wuttig and Riesenburg, “Sensitive Hadamard Transform Imaging Spectrometer with a simple MEMS”
Hadamard Transform
Benefits of Hadamard Multiplexing
• Decreased image acquisition time vs. scanning techniques
• Increased signal to noise ratio √n for Hadamard matrix (√n)/2 for S matrix
Wuttig and Riesenburg, “Sensitive Hadamard Transform Imaging Spectrometer”
What do we want?
• Focus a 25 mm square into a 10 mm square
• Bounce the square with specific pixels at specific angles (apply Hadamard matrix)
• Compress the light in one direction line
• Separate the line by wavelength
Goal: Capture the spectrum of the line
Why is this important?
• 71% mortality rate for diagnosed brain tumors
• Correlation between complete resectioning of tumors and improved prognosis
• Complete resectioning requires knowing the location of the tumor, especially tumor margins
• Imaging in a clinical setting should be fast