Discussion on GUI of Foundational Model Toolbox for Image Segmentation

[csking101]

Hi everyone,
I am Chinmaya, and I am interested in the "A graphical user interface of Foundational Model Toolbox for Image Segmentation" project. However, I have a few queries:

1. Should more emphasis be given to the GUI or to the support for the models in this project?
2. Is there an upper cap on how large a foundational model can be? (Because running the software locally and on a server are two very different scales)
3. Do we need to take into account memory limits? (Perhaps we can keep the foundation models cached once after download, and give an option to delete it as well)
4. What types of file formats should be supported? Also the if the dataset's resolutions don't match (for example, let's say you use U-Net), would it be appropriate to use techniques like interpolation?
5. How can we communicate about the required features in the software? (Because there will be a lot of them and it'd be better to graphically show the same)
6. For the UI/UX part, how much emphasis should be given on the design of the UI, should we make FIGMA models to start off with?

I hope the mentors can answer my queries/ guide me on how to get started!

Thanking you,
Chinmaya

chinmayasahu.211cs115@nitk.edu.in

[harshRaj2017]

Kindly let me know whether PyQt will be useful to learn for the GUI task or if StreamLit will be sufficient.

[A-SOLO]

Hello there,
I am Anupam Rai, student at IITBHU. I am interested in this (GUI of Foundational Model Toolbox for Image Segmentation) project.
I have extensively worked on image segmentation techniques. And I have experience with Python, TensorFlow, PyTorch, and PyQt5. Also, I have worked on projects integrating LLM and real-time LLM apps.
I look forward to collaborating and contributing to this project.

Thanking you,
Anupam Rai
anupamrai481@gmail.com

[austinhatch]

Hi Team,

My name is Austin Hatch and I am a software engineer from the United States. I recently got my Masters in Computer Science from the University of St. Andrews in Scotland. Since then, I have worked as the principal software engineer for a quantitative trading firm and the lead developer and founder on a web based ticketing platform. I am extremely interested in pivoting to building tools in the bioinformatics space and this project seems like an awesome way to contribute to the space.

I have extensive experience in full-stack application development, using a combination of Python, Typescript and Javascript (React) to build low latency, highly available web based trading applications, as well as desktop applications built using PyQt.

Through my Masters program I have experience in literature review, research and technical writing - as well as many facets of artificial intelligence and machine learning including deep learning with PyTorch. My work experience has allowed me to build highly functional and intuitive UIs on top of complex computational platforms.

I look forward to the opportunity to contribute my experience and knowledge in the space toward developing a UI for MRI Image classification, while being able to dedicate time to learning more about the space of medical image segmentation.

I would love to connect on this project and look forward to hearing from you.

Thanks,

Austin Hatch
ahatch.ai@gmail.com