Google Summer of Code @ INCF — 2018
Google Summer of Code is a well known program for students across the world to demonstrate their coding, leadership and collaboration skills. Every year a large number of organisations participate with various exciting projects. This year I got selected in International Neuroinformatics Coordinating Facility (INCF) as a GSOC intern. This blog is a brief on my experiences during these three months. This is my first ever blog, dont mind if I fail to make sense.
Pre — GSOC Period
Thanks to some friends and seniors at college for they made me aware about GSOC rules and announcement of organisations. Initially I was not very sure whether I should be applying or not for it takes quite a lot of time and effort in doing a proposal and then iterating on that. Though I gathered some motivation and selected some projects. I was most interested in Active Segmentation toolbox project at INCF because of two reasons :
- Project details were mentioned very appropriately on Trellis including the tasks to be addressed and set of deliverables. Mentors and organisation admin were very responsive, welcomed the newbies with every important information they wanted to know.
- It was a blend of Java based development and Machine Learning.
First I wrote a basic proposal and then did many revisions. I also got it reviewed from mentors once before the submission. This took a significant time so I ended up submitting only one proposal for GSOC’18. But somehow things worked for me and my proposal was selected.
GSOC — Period
Organisation had already made discussion threads for each selected student where he/she was expected to report weekly on tasks and challenges. My project was to develop a general-purpose interface that allows biologists and other domain experts to use state of the art techniques in Machine Learning to improve their image segmentation and classification results. During community bonding period I discussed the project structure, contribution guidelines and frequency of meetings with my mentors. During the coding period we first developed two more classification filters — Legendre Moments filter and Texture Descriptor filter and did a slight modification to previously developed Zernike Moments filter. Then we started integration of these filters and added the classification setting in the toolbox. We now have a working toolbox for both classification and segmentation task. We have already done some testing and look forward to make an official release very soon. We would encourage people to use the tool and report back to us with issues and suggestions. This is how the plugin looks like -
Some of these things were absolutely new to me which initially daunted me, but with course of time my endurance paid off and I managed to get through them. My work and everything else about the project can be found here.
Clearly this isn't the end of project and I look forward to contribute after GSOC period, fixing issues and improving on GUI. I believe that getting into this program was not that important as the learning that has come during this period. One more good thing is that you get very good mentors who not only guide you during this period and project, but may help you throughout the career.
To conclude, all I would like to say is that I couldn’t have a more promising, fulfilling and rewarding summer. I also take this opportunity to thank all the members of INCF community.