didsr / eedap Goto Github PK
View Code? Open in Web Editor NEWEvaluation Environment for Digital and Analog Pathology
Home Page: http://didsr.github.io/eeDAP/
License: Other
Evaluation Environment for Digital and Analog Pathology
Home Page: http://didsr.github.io/eeDAP/
License: Other
We need to update, edit and reorganize eeDAP manual.
We need to create documentation and possibly stand-alone executables that can do some of the following:
The current joystick speed is too fast to point on small feature (corner of a cell). We need this to do registration between camera and eyepiece. We should find a method to set joystick speed.
A new task to get WSI ROI position base on glass slide. It is only available in micro_rt mode
When doing the task across more than one slide, the image is usually out of focus. Auto-re-registration is not very good.
We might need a public WSI viewer for eeDAP. Then the users can enjoy the features in it (e.g. annotation) and we don't need to worry about color problem.
Recording the stage position continuously can help a lot to track pathologist reading path.
I clicked on Refine Registration. High-resolution registration began. The WSI patch was shown, but the stage did not move to the corresponding location. Of course, I could click "Load last calibration", but it should be done automatically.
If you want to be added to the mailing list, simply submit a comment to this
issue asking to be added.
Original issue reported on code.google.com by [email protected]
on 20 Feb 2014 at 7:05
We may need to try different channels or contrast enhancement.
This issue is in response to email from Qi. Qi wrote that: "The registration of PHH3 during task part has a little problem. there are about 3 out of 40 images can't register even with best_register button. We never face this problem for HE image."
Thanks with your email. Yes, let's create a new version v4p0, including the compiled version for windows.
I'm rewriting your summary here for you to describe the updates. Please add the model numbers:
@qigongFDA said ...
The new task name is task_count2_imagescope.m. (We use
exportXML_ROIandCircle.m to generate .xml file).
There are 2 new sample input files to test all tasks. Please choose
them base on the system
a. tissue40x-8B-axioplan2i-ludl-allTasks.dapsi
b. tissue40x-8B-olympus-ludl-allTasks.dapsi
There is white balance problem for camera (camera image looks green or red). We need to improve the white balance to fix this problem. It can improve the registration accuracy.
The location I selected was near the boundary and I got the error below. The process shouldn't cause an error. I should be allowed to select another registration location, not so close to the boundary.
Error in ExtractROI: Illegal Extraction Origin Left or Top, ExtractHeight or ExtractWidth
Origin Left = -113.2128
Origin Top = 11059.5154
InputWidth = 1654.0152
InputHeight = 1654.0152
OutputWidth = 300
OutputHeight = 300
Video button should reopen camera image.
I will show Qi the reticles we have. We should catalog them and draw them in FDASVN eeDAP\000_docs\eeDAP_screen_shots_and_drawings\reticle.pptx
Qi, you need this for the eeDAP-version of the mitotic counting study.
Jessica could possibly do the drawing after watching Qi do one.
Brandon
Is there any capability to handle jpeg2000?
Save the exe somewhere on the research cluster
We need to add a new button to Hide/display reticle on WSI image. Because sometime reticle might block cell and affect study result.
Please create a tool on NCIPhub that runs eeDAP.
Create a way to show the basic information in an image and to test if bioformats can open it and then show a thumbnail version.
The original eeDAP software had pan and zoom capabilities. We need to reinstate these features.
When I click "Take Stage Position 1" (or 2 or 3), the button stays active until the image acquisition from the camera is complete. It should become inactive immediately to avoid double clicks.
Same with Take WSI Position 1 2 or 3.
We need to add a new button to skip registration between eyepiece and camera, after high resolution registration. Then for the same system, user doesn't have to do eyepiece and camera registration every time.
This is in response to Qi's email: "The response of new stage MAC 6000 is a little slow. We send moving comment, it has a few second delay before moving.
I think the delay comes both from MCR and MAC6000 device. I didn't try to fit it today. Because it is working now, and the delay is not very long. I am afraid of brake it during the fixing. You can have a look of it next Monday. If you want to improve it, I will try to fix before next friday."
Create features to record camera video and user voice.
I shouldn't have to click reset and then fast registration
Hello everyone,
I think there might be a easier method.
We can edit our code to only generate xml file but not show WSI FOV. Then compile and install it on NCIPhub. Similar as our previous APP one NCIPhub, we can download and xml and output files from NCIPhub. And then we use imagescope to show the WSI with FOV annotations. In this case, we don't have to upload image to NCIPhub, and also don't need to worry about color. There might be a problem that it is a little hard to open imagescope automatically by NCIPhub.
Best wish from China,
Qi
On Friday, August 25, 2017 12:01 PM, "Gallas, Brandon D." wrote:
Yes, NCIPhub is the desired place for developing and hosting this tool in the cloud. It’s FREE.
I am adding Qi to this thread. He is in China for another two-six weeks. He may have a delay to his trip because of his visa. We are delaying the trip to Memorial Sloan Kettering and the reader study until October. That said, it will still be nice if we can run this in the cloud sooner to test and maybe even collect data in digital mode.
Qi created a tool for the Matlab gui that I showed both of you. I think the NCIPhub server has a licensed version of Matlab.
If they grant us space to upload big images, we are in business for digital mode with two shortcomings:
• We won’t have a special feature, opening a third-party app because it only runs on Windows.
• There will be some questions for us to deal with down the road regarding color.
Have a good weekend.
Brandon
From: Berman, Benjamin *
Sent: Friday, August 25, 2017 12:59 PM
To: Gallas, Brandon D.; purva sawant
Subject: RE: cloud version of eeDAP
Hi Brandon and Purva,
Briefly looking into running eeDAP from a container. As far as Docker compatibility is concerned, it seems there would be two directions:
Compile eeDAP using Matlab Compiler (a Matlab product). Then install the Matlab Compiler Runtime (MCR) into the container along with eeDAP. From my understanding, this ends up being an overly large container (MCR is designed to handle all of Matlab’s capabilities).
Generate eeDAP using Matlab Coder (another Matlab product) as C and C++ code. Then use a C compiler, and load the compiled eeDAP into the container. Could be doable, and much more portable, but I am sure there will be plenty of debugging to do.
Interesting to find on a few stack-overflow exchanges that Matlab developers are getting onboard this contained, portable code idea. I wouldn’t be surprised if we see yet another Matlab product within the next few years to help achieve this. Attended an NCI meeting on quantitative imaging a few months ago and everyone was talking about Docker.
Just wanted to add my 2 cents here. I haven’t looked into the NCIP hub, but I know that Brandon and Qi (when he gets back) have plenty of experience with that.
Have a nice weekend
-Ben
From: Gallas, Brandon D.
Sent: Thursday, August 24, 2017 5:30 PM
To: purva sawant
Cc: Berman, Benjamin *
Subject: RE: cloud version of eeDAP
Purva, Thanks for talking with me today.
I think we agreed that you would look into a plan to install eeDAP in a container that would serve the program from the cloud. While you are working with Amazon for Aldo’s project, I’m interested in seeing if the NCIP hub would work for us. It’s free and they are quite friendly since they are government (NCI) and they want more users on their site. They have always honored my requests for more space and often email me to urge me to create tools. Please start there.
Below are links to the eeDAP source and compiled app, documentation for producing tools at NCIPhub, and the NCIPhub main page. Notice that the documentation is at hubzero. NCIPhub is a specific server of that software hosted by the NCI.
https://github.com/DIDSR/eeDAP
https://hubzero.org/documentation/2.1.0/users/tools
https://nciphub.org/
https://nciphub.org/resources/tools
For now, I’d like to know how long you think the development of this might take. If possible I would like to use it for a study Sept 11.
Let me know if you have any questions. Also Ben Berman is on the cc line. He has run eeDAP and can sometimes help with that.
Brandon
The program should not fall into an error mode. It should display a message and then return without issue.
Can you do the following?
imageScopeXMLcreate: input parameter: fileName, creates an empty xml file for fileName
imageScopeXMLaddCircle: input parameters: filename, circle radius and x0,y0 pixel values at center
imageScopeXMLaddRectangle: input parameters: filename, xLength, yLength and x0,y0 pixel values at center
imageScopeXMLaddCounter: input parameters: filename, x, y
What do you think about adding a button to the data collection gui that allows the user to do the eyepiece/camera registration?
I think we should activate the next button sooner.
See if you can show the same FOVs.
It should be a simple scale factor between the location and size for their image (.23mpp) vs our image (.25mpp).
We should move items from FDA SVN to GITHUB
The plan is to design a feature study to compare two different displays for use in digital pathology. The study is in the spirit of the mitotic counting study described in Brandon's 2016 Pathology Informatics Summit presentation. The NHS team will consider the task to be performed. Here's some characteristics for the task.
Please refer to the front page of the eeDAP GitHub project: https://github.com/DIDSR/eeDAP
The README on the front page contains links to the manual and software releases, as well as related references and resources. The front page also contains the top-level directory of the source code repository. Check it out.
Videos are created and saved in /raidb/qigong/eeDAPvideoManual.
The digital mode script is saved in the same directory.
I have ImageScope.exe in my bin directory.
I ran “wine ImageScope.exe”/
I tog the following error:
err:module:import_dll Library MSVBVM60.DLL (which is needed by L"Z: \home\Formula1\Desktop\ImageScope.exe") not found.
I found someone got this to work with “winetricks”
https://ubuntuforums.org/showthread.php?t=2144183
Could you give it a try when you don’t want to do your own work?
We need let our new stage ProScan III working
Please include an "Update Summary" file like you did for iMRMC discussing the major updates.
We never used ROI_W, ROI_H, MoveFlag, ZoomFlag Q_Op1, Q_Op2, Q_Op4, Q_Op4 for most tasks input. These item will confuse end user. We will remove them from all current tasks.
Get new USB 3.0 camera work.
Upload eeDAP to NCIP hub could. Then user can use it without installation.
If the file name is incorrect, the program closes with errors. It should close more purposefully.
We talked about errors that appeared when running eeDAP with a new version of ImageScope.
I believe the errors are from ImageScope has a new format for the xml file.
It is not a big difference because everything will still run after clicking ok to the errors.
Please get to the bottom of this and let me know.
Brandon
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.