Sue Lee, Week #6, Finish Line

My lab during an ice cream outing at Toscanini's :)

This week, I proceeded to attempt converting a raw SEM image into a binary array. Although the code successfully segmented the images and calculated the centroid coordinate points, the order in which the coordinates of the centroids were listed in the data table was incorrect. Upon research, I realized that the root of the problem was the method in which MATLAB’s ‘centroid’ function saves the coordinate points of the centroids. This led to incorrect calculations of the state values, and thus an inaccurate binary array. To overcome this problem, I manually rearranged the coordinate points so that the y-coordinates would be in the correct order of descending posts from top to bottom (rows) and left to right (columns). As a result, the code produced a binary array with the correct numerical values, as well as a visual representation identical to the block copolymer pattern shown in the raw SEM image.

Although the code proved to be relatively successful in identifying the presence or lack thereof of a connection between each pair of adjacent posts of a given block copolymer pattern, it remained to have imperfections like the centroid function that ultimately required manual manipulation within the code. Furthermore, the SEM image used was an ideal image with high contrast and a significant distinction between the bright posts and the black background, with a simple pattern created on single, equally spaced posts. Hence, the code will need further modifications to be usable for more impure SEM images. Despite this, overall, the code demonstrated proficiency in being able to accommodate itself to different forms of images and convert them into binary arrays that accurately represent the connections present within a block copolymer pattern.

I wrote a six-page report on what I did during my time here, which I handed to my professor during my final meeting with him on my last day. He gave me some advice in going forward in my life in terms of college and grad school, and encouraged me to continue pursue programming. He told me that unlike physics, what he majored and absolutely loved, computer science was literally limitless in its capacities. Nowadays, he said, there are many computer science courses that are integrated with other subjects like economics, biology, and materials science, which relates the real world to the rather abstract world of code. I've always loved both humanities and science, and it was inspiring to know that there are truly more ways to integrate both worlds in my studies as I move onward.

My time at Dr. Berggren's lab has gone by in a second, and I definitely wish I had planned to stay here longer. I will miss all the talks with my grad student, my professor, and the rest of my lab members, as well as the corny jokes in the "lab snack policy" emails (we now have a high-tech dispenser for nuts and granola). All jokes aside, I have learned many invaluable lessons throughout my time here. I learned to push myself and not be complacent with a code that can be, in fact, improved to a limitless capacity. I learned that no code is ever perfect, and the only way to come even fairly close to perfection is through endless revision.

Thank you to everyone who has made this experience possible.