Fit (One) GMM to simulated, 2D dataClean up latex code on overleafMerge PPN2V with PN2VBook ticket to SSMeeting With MP
Book ticket to LReply to KG and DSMeet FJMeeting with MB
Meeting with AGCheck KK Data
Reply to MHTSend out vacation emailGeneralize above to n GaussiansReply to KG
Generalize GenSeg to n SimulatedObjects
Generalize GenSeg to n Actual Objects
- Get Threshold from Pixel Intensities
- Use fixed threshold and see GMM performance
- Look for QIP example and solver
- Test on MHT's Graph Strcuture
- Label membrane data
- label fixed specimen
- Solve the 512 * 512 issue on AK's data
- Test segmentation with StarDist
- Perform Multiview Fusion on JG' data
- Produce quadratic matching on two time points separated by few time points
- Perform experiments with HydraNet on Dataset1
- Perform experiments with HydraNet on Dataset2
- Implement NMS on MultipleGaussian
- Check KK Data from the perspective of graph matching
- Move data to the correct location
- Find the right parameter for bleach correction
- Try making the pipeline for 2D data
- Test Gurobi (Quadratic Matching) on Mastodon, Platynereis
- Extend MultiViewFusion
- Perform Segmentation on MHT Data
- Check performance of segmentation with MHT annotation
- Write ILP tracker with Mass Conservation on MHT Data
- Perform Tracking on MHT data
- Check performance of Tracking on MHT Data
- Document performance
- Push the article to arxiv
- Create validation results with StarDist 2D on data1, data2
- Test GenSeg on Actual Data
- Perform FairSim on VY Data
- Perform Segmentation of Myson and Actin with Pytorch
- Extend Pytorch for blob detection
- Document and communicate performance to VY, AB
- Perform StarDist on JF Data
- Document and Communicate to JF
Prepare denoising pipeline with PPN2V, Bootstrapped for BV- Add Care notebooks
- git rebase the PPN2V repository
- book L, UK hotel
- read Jetley et al paper
- put the Genseg code into a class
- How can StarDist be used for a three-class detection?
Create fileserver for KM's dataCreate zip file of K's DataCopy zip file to correct locationApply for project space increaseUnzip the contentsReply to KMGet documents for A33Fill up the form of A33File the house cancellationMeet IOCopy the data from RL's doc to excel
- Make a new branch for simulated data
Scan the ID documentPrepare envelope- Implement the 'Learning Graph Matching' for C. Elegans dataset
- Explain the theory of Learning Graph Matching in post
- Start looking at KM/MB data
- Prepare simulated data three quality levels
Prepare presentation to discuss with PAvel- ~~Figure out why you are unable to copy data into fileserver (uid=lalit) ~~
Write a macro- Perform convolutions using pytorch
- Prepare and save simulated data
- Run the first analysis
- Push to Github
- Create a new branch for simulated data: GenSeg
- Maybe write a post about it
- Read up on differentiable IOU between masks and distribution
- Make up a new branch for differentiable IOU
- Read up on N2V SimSim (what could be the next step)
- Start summarizing the data from JF
- Start summarizing the data from KM, MB. Perform analyses
- Start summarizing the paper1, paper2, paper3
- Book hotel for QBI
- Start looking at the biological data
- Test your learned model on this data.
- Clean up the code
See AK's SlidesSee MP's Slides- Document Style Transfer GAN
- Check out FJ's threshold idea (sum up intensities for each label, look at 10 thresholds and choose)
- Use network to predict GMM parameters
- Prepare Slides
Document things which AK mentioned yesterday- Write letter to Lund University folks.
- Finish Analysis on Keller Lab
Pay for kuche- Figure out if starpose makes sense. (Typically you have images and not masks)
- Read DeepPose.
- Read Gatys et al.
- Document Gatys et al.
- Try Gatys et al on Simulated Data.
- Document Keller lab Analysis on Simulated Data.
- Try coloring pixels with different colors
- Perform unsupervised matching of sub-lineages.
- Complete GenSeg.
- See if you can perform N2V on optical flow.
meet KK.- Figure out the cluster.
- Modify the markdown timetable such that it can print specific months.
- Answer the pull request.
- Document KellerLab.
Document TextureSynthesis.- Train a Unet baseline on DSB and save.
- Finish transforming pytorch pn2v for segmentation.
- Train pn2v baseline on DSB and save.
- Add second post for documentation on texture synthesis.
- Add Mouse Actin Notebooks.
- Add Mouse Skull Nuclei Notebooks.
- Add ReadMe.
- Document DCGAN.