Elias Nehme
@EliasNehme9
Followers
38
Following
38
Statuses
14
PhD Candidate at Electrical Engineering working on Computational Imaging and Machine Learning.
Technion-IIT, Haifa, Israel
Joined February 2021
RT @guy__ohayon: Posterior sampling for solving inverse problems? No thanks😉 Check out our new photo-realistic image restoration algorithm…
0
8
0
Check out our recent collaboration at @ShechtmanLab with @SartoriusGlobal! As we show in our new @NatureComms paper: the depth-sensing capabilities of high throughput microscopes can be significantly enhanced using "clever" compact optics & deep learning
This enables PSF engineering in space-limited microscopy systems, e.g. that fit into incubators. Uses include extended depth-of-field and snapshot 3D spheroid from 2D, using algorithms by @EliasNehme9. Bonus: Obtaining training images is easy with a high-throughput microscope.
0
3
11
RT @TamarRottShaham: Accepted to #ICML2024!! 🚀 Meet MAIA- a Multimodal Automated Interpretabilty Agent that helps users understand AI syst…
0
49
0
RT @_akhaliq: Blended-NeRF: Zero-Shot Object Generation and Blending in Existing Neural Radiance Fields paper page:
0
66
0
@sfhbarnett @LaineBioImaging @ShechtmanLab This seems like a problem with the construction of the training set. It could be a bug somewhere in the Colab, but I think the crops are probably not simulated properly in figure2, hence the weird behavior. Could you please share your notebook (via email) so I can take a look?
1
0
1
@sfhbarnett @LaineBioImaging @ShechtmanLab Of course, the previous statement is true assuming neither is overfitting. As for the other point - yes I meant the Batch Norm layers
1
0
2
@sfhbarnett @LaineBioImaging @ShechtmanLab As for the x8 gridding artifact, I only encountered it when the training data and the testing data had a distribution shift in terms of either SNR, density, or PSF mismatch. I think this probably stems from the U-net 3x encoder-decoder (2^3=8) with incorrect normalizations
1
0
2
0
0
1
RT @docmilanfar: Tweedie's formula is a rare gem💎 Setup: measurement 𝐱 of quantity 𝐮 𝐱 = 𝐮 + 𝐞 with 𝐞 ∼ 𝒩(𝟎, σ² 𝐈) Find: min mean² esti…
0
45
0