Category: SAS / Image Analytics – Kennedy S

Today I managed to get Biomedical images to work in SAS studio. The one thing though is that I can not upload any file over 100mb of data. I started to work in SAS studio with biomedical images. Basically now I am using SAS studio and combining that with everything else I did. So instead of masking something with ImageJ I now do that in SAS studio. I did run into one issue where the image just went full black. However that somehow resolved itself, but when trying to do another image the same thing happened. I met with my two supervisors who helped me with the problem. I also met with my overall supervisor and just had a quick check in and discussed everything I had learned so far.

Today I found out that I can not use Jupyter Notebook with CAS, this is because of the way I had to launch it. Instead, I now am switching to SAS Studio and SAS’s coding language in order to work. The reason I would like a CAS server is that it basically speeds up processing as my computer is no longer the computer that is doing the processing. I then proceeded to mess around with some of the examples in SAS studio however could not get biomedical images to work, I am currently working on that. One of my supervisors also sent me code examples for biomedical images that I can try. I am still working on getting biomedical images to work.

Today I installed SWAT and started to get connected to SAS’s CAS. This is a system that will allow me to integrate my python code with SAS Viya and also work with the computer visions team. I also used more examples from ITK in Jupyter notebook

Today I got thresholding to work with Jupyter notebook. This allows me to replace pixels of a certain color with another color such as black and white. I also encountered an issue with ImageJ however I was able to contact my employers and get that sorted out.

Today I got Jupyter Notebook working. This program will allow me to build pipelines. I mainly messed around with the smoothing code which smooths an image’s edges. I also experimented with Threshold. Tomorrow I will continue manipulating pixels and seeing what else is possible. I also had and or have to modify the code used as some of it is not needed for what I am doing.

Today I worked with Aliza software. This software can convert CT scans to 3d models, you can also overlay multiple scans in order to see different data. In the video below you can see a 3d model of lungs that are infected with COVID-19. Later in the video, the black blobs represent infected parts of the lungs. Also today I worked on getting Jupyter Notebook working. This should allow me to start manipulating the CT scans.

This morning I met with two employees of the SAS Computer Visions team. After the meeting, I worked on getting the software up and running. You can see an example below with ImageJ. In the example, I masked the lungs of the CT scan. After lunch, I continued to work on getting the software set up however I ran into some hardware and software issues. I am currently working to resolve them.

For 8 days I will be shadowing and working with SAS’s computer vision team. They work on multiple different projects. I will be working specifically on analyzing 2d and 3d biomedical images