We are going to try something new that we introduced in 2020. This is in addition to your blog post. The goal is to catch things early if you are behind in your timeline early.
Well, I am not sure if we have to post during this week, but I would like to share my progress.
What progress have you made this week?
I have found the required datasets on Kaggle and / or requested access to data from the owners and uploaded it to my Kaggle profile
What do I plan to do next week? After the meeting with my mentor I will come up with a plan, but for now I will just continue studying papers related to my work
Have you had any blockers or issues that are impeding your project? The only problem is the large Cholec-80 dataset, which is hard to process, but I guess that using stratified samples for training and a fixed sample for validation is fine
What progress have you made this week?
I first finished with the coding pattern and architecture planning a day before coding period started. Then I made the following progress:
Created the traditional login flow with username, password and serverURL. There was a slight modification from the ECEB app as the organisation doesn’t have to be manually entered. MR can be seen here.
Created local authentication for storing credentials for fast logins so medical professionals working in different organisations don’t waste time in tedious effort of entering things manually all the time. MR can be seen here.
Created a basic CI/CD pipeline. MR can be seen here.
Setup localization of the app.
What I plan to do next week?
Next week, I plan to implement addition of TEI in the homepage and create a mock peripheral. Then initiate connection between mock BLE and the app.
Have you had any blockers or issues that are impeding your project?
@sunbiz I needed some information about initial data needed to be entered while enrolling a baby. I have read up some things in that matter but it will definitely need some improvement.
Mock data vitals which were given from the device during testing phase are also required.
These things will be needed in the subsequent week.
I have been learning FHIR and did the setup for HAPI’s FHIR Server and started the React project
Forked the repo and removed the angular app and created a React.js project in it with the typescript template.
Got a good understanding of the FHIR Standard and wrote this blog
Decided on a data fetching architecture using React Query and Axios (the company I am currently interning at uses this, and it is highly scalable)
What progress have you made this week?
Next week:
Implement the routes
Add the data fetching architecture (which involves caching) to the project
Set up the CI and CD pipeline for the project
Implement the PatientCRUD using this architecture
Have you had any blockers or issues that are impeding your project?
I had some issues with the HAPI FHIR which are now resolved (was related to JDK version on my machine)
P.S. The work I do is saturated during the end of the week, so Friday updates for this week don’t have the work that I think I can do in a week, so next Friday the work will be more than I do in a 5 day week it will instead be the report of a full 7 day week.
Found out a stable way to “simulate” VR for classification and segmentation task
Found an interesting approach to object detection on distorted images
Found a thread where a way to distort points by coordinates is discussed. It can be applied to object detection
What do I plan to do next week?
Try to use the distortion tool for image segmentation
Search for any other papers / data / related work
Experiment with distortion bounding boxes
Have you had any blockers or issues that are impeding your project?
The main problem is the lack of open-source data. The data is someone’s “know how” or simply not published. I guess that it is easier to find a way to distort points rather than to search for data
Like we discussed, we have requested piglet data from our collaborators at Purdue and will share with you to add to the app as demo/mock data.
@all students - please create separate issues on which you are working or planning to work on. Please submit your MR with reference to the issue. We want you all to learn about working in an open-source environment, as much as we want you to finish the project at hand.
I added Material UI and implemented the Layout with Mobile Responsive Navbar, React Router, and Dynamic Breadcrumbs MR
Added a CI CD pipeline for the app, which is deployed on Netlify and has two deployments, Preview and Production
Added the React Query + Axios Data Fetching Architecture (Caching works out of the box) MR
Implemented the get and search patient using this architecture MR
What progress have you made this week?
Next week’s main focus would be to integrate lh-toolkit-webcomponents into my project anyhow and then finish then Patient CRUD
Next week:
Implement the Patient Dashboard UI
Add create and update in the Patient Flow
Find an FHIR API that supports Radiology and implement Add Radiology Order
Have you had any blockers or issues that are impeding your project?
I have been having issues when installing lh-toolkit-webcomponents into my project, I have updated the package.json of this repo in hopes of those issues getting fixed. MR
HAPI FHIR’s test API doesn’t support Radiology and it is the next thing that is on my task list (following my GSoC proposal) If I don’t find an API that supports it this week I will move forward with other parts of the project. @sunbiz please do suggest some APIs that support Radiology forum post
Developed multi class and binary segmentation models.
Used the proposed technique to adapt their domain to AR
What do I plan to do next week?
Do a comprehensive analysis of results, search for improvements
Probably start the classification task
Have you had any blockers or issues that are impeding your project?
Adaptation of a binary model was simple, however in case of the multi class one I had to change the pipeline a little. Also, the resulting performance of adapted models is lower that the performance of the original one
Next week’s main focus would be to discuss the project’s future with @sunbiz and start with the next thing in the pipeline
[[ Will update this section after the meet with sunbiz ]]
Have you had any blockers or issues that are impeding your project?
Next week’s plan is to first set up a local HAPI FHIR server and then, scale the search functionality of the patient which will tie into searching for Appointments as well. Once the Search in Appointments is done, I am planning to integrate one in Patient Dashboard and implement CRUD for Appointments
Have you had any blockers or issues that are impeding your project?
My University started this week so got acquired in assignments, and I will be acquired in assignments this weekend too.
The test.hapi.fhir server is not very reliable, I have seen 2 downtimes since I started working on this project. I reached out to the FHIR community and saw there was a backup server but it was blocked for me because of CORS
So, I am planning to set up a local HAPI FHIR server so that downtimes don’t affect this project’s development.
Finished adding TEI functionality along with the avatar.
Avatar functionality merge request.
Made UI improvements in the flow.
Connected mock ble device for temporary purpose to the required TE in the app.
What I plan to do next week?
Using a mock BLE to display the data over this weekend.
Finish minor functionalities like clear cache etc.
Write functions for data sync in hive storage repository that will be used later but are related to the current work.
Have you had any blockers or issues that are impeding the project?
On experimenting with the temporary mock ble for a good amount of time, I observed there are some lags and glitches that are happening. To create the mock BLE I used code based on the bloc library source code i.e. from scratch implementation of bloc. Bloc itself has been considered to be a heavy architecture. Will probably replace it with getx or a better alternative.
Switched to Oculus Integration SDK for Hand Tracking
What do I plan to do next week?
Complete the Oculus Audio Recording GUI
Correct Hand Tracking Issue
Have you had any blockers or issues that are impeding your project?
Switching the SDK caused a lot of issues as OpenVR backend is used, whereas in XR integration toolkit OpenVR is used in only few components. Corrected most of them.