Page 389 - AWSAR 2.0

the scanner to complete the scan faster. We then have to find a way to estimate the missing data to reconstruct a proper-looking image devoid of any artifacts. Research in this area has been going on for many decades now. But a new technique called compressed sensing recently became so popular that its application to MRI reconstruction was of huge interest to the community. There laid my chance to make a meaningful contribution and earn a Ph.D.
During the final year of my postgraduation, I had already developed a huge interest in further understanding the process of image formation in general and its various manipulation methods. So, this was just the right thing that could have happened to me, and I truly believe that it had happened for good. I, wholeheartedly, grabbed the opportunity and started working on it right away. Much like the journey of any other researcher, mine also started with understanding the literature, studying all the different ways in
which people have already tried
to reconstruct the magnetic
resonance (MR) image. The
first task my guide assigned
was to identify a missing link/
knowledge gap in the literature,
where I could make an original
contribution. I was convinced
that the better one understands
the literature, the closer one will
get to find that gap.
Mr. Ajin Joy || 365
We noticed that almost everywhere in the literature, when edge preservation properties are required, a method called ‘Total Variation (TV) Minimization’ is used. Well, there is a saying that too much of anything is bad. So, I started feeling a bit skeptical about it. The more I discussed the literature with my guide, the more the feeling grew. Since edge is not the only feature in MRI that needs to be preserved, TV was often used along with other methods that focus on preserving the smoothness of the image as well. However, these other methods continued to evolve, to the point where it was capable of delivering much better stand-alone reconstruction compared to TV. Therefore, it became clear to us that TV was holding some of the performance back.
We tried to present a solution in the form of Perona-Malik (PM) diffusion as an alternative to TV, which is also capable of edge-preservation, with an added advantage that one can also
control the strength of edges that are preserved. The results looked quite promising. “This should clearly be it,” I told myself with a sense of achievement. But the joy only lasted till I realized the fact that it was one thing to find a better method of reconstruction, and it was another thing to convince a journal to publish it. It became clear that when it comes to publishing the work, one has to master the art of presenting it as well. After a few rejections, I wondered what more could I have done. It turned out to be a tough task to convince the reviewers that PM diffusion is a
better alternative to TV.
Fortunately, I have never stopped
believing that everything happens for good, and I never felt short of encouragements and
MRI scanning rarely needs any introduction. Over the past five decades, the technology involved in it has really matured to make it a widely used imaging modality to picture internal organs of the human body. It has helped numerous people by aiding in early detection and staging of numerous diseases and monitoring the body’s response to various treatments.
Soon, much to my
delight, I found it in the area
of edge-preservation based reconstruction approaches
in MR image reconstruction.
Edges can be imagined as
the boundaries separating
different regions in an MR
is straightforward to understand that its preservation can become crucial for a good diagnostic quality of the reconstructed image.
image. So, it