Ja myös #Facebook-muisto unesta samoin kuuden vuoden takaa: ”Unessa kävin junassa, jossa jonkin #aikadilataatio-ilmiön takia äänen nopeus oli muuttunut. Ihmiset yrittivät puhua h-i-t-a-a-s-t-i mutta lähes kaikilla pitkät ja lyhyet äänteet menivät päinvastoin. Minä asiaa tutkineena osasin paremmin, kävi ilmi junassa tehtyjä äänitteitä toistettaessa, mutta minäkään en saanut onnistumaan luontevaa puhetempoa.” #unet

Extensibility Of U-Net Neural Network Model For Hydrographic Feature Extraction And Implications For Hydrologic Modeling
--
https://doi.org/10.3390/rs13122368 <-- shared paper
--
#GIS #spatial #mapping #machinelearning #neuralnetwork #Unet #featureextraction #hydrography #hydrology #water #automation #remotesensing #surfacewater #NHD #bluelines #topography #deeplearning #AI #artificialintelligence #LiDAR #ifSAR #elevation #topography #ANN #HPC #model #modeling #trainingdata #Alaska #USA #watersheds #DTM #DSM #DEM #routing #network #D8 #maps #opendata #3dep #elevation #topography #USGS

My first Kaggle notebook on semantic segmentation with U-Net
https://www.kaggle.com/code/muhammadwasee/u-net-segmentation

It is also available on GitHub
https://github.com/hwaseem04/unet along with notes for the U-Net paper

Special thanks to @SebRaschka
for his amazing ML with PyTorch book for building my PyTorch foundations

"CartoCell, a high-throughput pipeline for accurate 3D image analysis, unveils cell morphology patterns in epithelial cysts" by Andrés-San Román et al. 2023 from @lmescu's lab

Learns from high-resolution images how to segment well on low-resolution imagery–hence the high-throughput.

Applies a 3D ResU-Net network to transform fluorescence images into segmented cells. And blows the competition out of the Petri dish!