Quote from elammertse on June 10, 2021, 8:00 pm

Hello all!

It was great to meet you all at the get-together workshop on June 8th. Special thanks to Massi and Rune for putting it together; it was a great introduction to defocustracker itself and the vision for the project moving forward. 

While I have not yet used defocustracker, I am currently developing a parallel approach in the same paradigm and evaluating deep learning models to perform the Z-prediction tasks in comparison to normalized cross-correlation. Our end goals are the same, to provide a widely accessible 3D tracking tool, and the effort put into defocustracker and  this website to achieve that goal is impressive. I also appreciate the open-source philosophy of the project, and see clear opportunities for collaborative development, dataset sharing, and establishing a common framework for performance evaluation of new tracking algorithms.

My immediate reactions to our discussions in Tuesday's workshops are:

1. I need to try out defocustracker! I have a large training dataset of brightfield particle images and several corresponding experimental datasets, and I am very curious to try to incorporate them into defocustracker's framework and evaluate the performance. Getting my feet wet there may provide some ideas for future development of the application.
2. I am very interesting in contributing to the development of defocustracker in the open-source framework. Like most (all?) of you, I am not a software developer, but as a later-stage PhD student starting the science/engineering job search process, I am eager for opportunities to build some development skills, especially in a collaborative environment. That would be great on its own; doubly so if the project aligns well with my experience and academic interests. I have intermediate, but by no means expert, experience with both python and MATLAB and would appreciate the opportunity to move from ad-hoc scripting to contributing to a project of a wider scope. 
3. Given (2), my next question is, "Where do I start?" I asked in the get-together about what high-level plans, or prioritized feature "wish lists", may exist for the near to medium term development of defocustracker. It didn't sound like anything that concrete yet exists, but that would be a perfect use of this forum, to collate ideas and wishes from the user base together to drive some initial development effort. 
4. "Where do I start?" quickly becomes "How do I start?" My immediate tasks are to familiarize myself with defocustracker, git, and GitLab, and learning the actual mechanics of the code and how to go about creating a branch and tinkering around. So, I hope to be able to focus a bit on that in the coming weeks. (In between finishing this @#$@# manuscript....)
5. My interests right now are in interfacing our datasets with defocustracker, and seeing how I could implement our Z-prediction algorithms within the defocustracker framework. Beyond that, it may be interesting to implement more advanced particle linking algorithms, or additional Z-prediction algorithms I plan to explore, but I need to get my feet wet first (1 and 4 above)...  

Anyways, that's where my head is currently at. I hope to share some additional thoughts as I begin to play around. But for now, again, it was great to meet you all! Looking forward to additional workshops that may be put on in the future.

Thanks!

Evan