new search line

correction on the lbfgs method init

introduced LBFGS and frequency regularization

train and test now working properly

introduced lambda for maxima number optimization

git is a baaaaad

Merge branch 'master' of ssh://c4science.ch/source/EB_ECG_Smart_Resampler

Merge branch 'master' of ssh://c4science.ch/source/EB_ECG_Smart_Resampler

FML

optimizing frequency transform statically After level definition through…

FML

implemented spectral learning

First experiments with multy sinc interpolation

IDFK

error in size allignment of t_base

Added support for larger dataset generation

settings for DESKTOP at work

Merge branch 'master' of ssh://c4science.ch/diffusion/12761/lc-sampling-theory

change in sinc interpolation standard

implemented vmap instead of list comprehension

removed unrelated files, Implemented un-batched version on the gaussian…

removed pycache

added README

added dataset folder creation

Updated gitignore, removed the dataset

initial commit

added gitignore

removed jax test

minor corrections

Merge branch 'master' of ssh://c4science.ch/source/EB_ECG_Smart_Resampler

compute srf test

Tidied the source directory, Added atwo custom scripts for templates…

corrected typos in the final results computation

Added general results computation for each level

Modified printing statements to be clearer

found and solved a condition that stalled the delineation

decresed the maximum time given to ecgpuwave to complete from 20 to 7: this…

Improved segment warping: solving a bug that wrongly overwrote some segments…

now the time-weight in ddtw is parametrized and can be used at a param at launch

rolling back to a moderate value (last test wrongly executed)

incresed time lassoo in ddtw_tv to aggressive values so to check the effect…

increased weight on timing informations in ddtw_tv

changed timing weight in ddtw_tv

increased the skip timing to 300 seconds if the delinaetor fails

implemented differential event based dtw

improved previous design still using dtw standard

introduced delineation, snd dtw distance results. Started working on a new…

introduced an "improved" version of wfdb.csv2mit. Before, the changes where…

keeping csv data: analyze the wfdb behaviour after

correction on saturation value. delineator still missbehaving

chenged timout value for ecgpuwave to accomodate slower executions

temporary patch to obtain results: ECGPUwave is timeouted after 20 seconds to…

typo

improved messages clarity

changed dataset extraction. moved from RMS to PRD

added timing mode informations

updated gitignore to ignore the results folder Redirected output to a custom…

implemented parellilization on the levels

Implemented multiple sre-sampling methods composite evaluation

improved the ADC definition, implemented multiple resampling strategies

implemented 0-ing at start and stop of beats in precise computation, solved…

added precise implementation of EB-sampler

minor semantical improvements. Started working on the exact algorithm that…

lowered median filter size

changed timing in median filtering

modified saved data structure and data handling so to improve memory efficiency

changed implementation. Now every level is computed independently. still memory…

solved minor bug that made the median filtering not working properly (due to…

changed saving method in saving function from custom to pickle. Implemented 0…

improved the lvl crossing algo. Templates estraction now work heyrarcically.

imploemented old templates check for inclusion in newly founded clusters

imported and implemented the dinamic time warping distance with a c core.

Solved minor bug in template selection

implemented harson-darlinghton tes and improved

paramterized the percentage of beats per cluster (argument)

the percentage of points per cluster is now parametric

modified the percentage of sampes needed to keep a cluster

improved templates creation and added logs for time every tramplate was used…

the saved templates from the progressive template

updated with cleare names

new method for template recomputation

Created a "one shot" multiple template analysis which compute a number of…

modified the nemae of some source file with more descriptive and short names.

minor bug solution, implemented different type of template

implemented segmentation based template reconstr

implemented naive template reconstruction. Started implementing segmentation…

It is now possible to build the ECG beat template form a file using "build…

better README

Started investigating the application of DTW to ECG beats. Improved README

minor bug in import solved

beats extraction and saving of signle beats and generic pre-processing routin…

pruned useless scripts, moved helper script in appropriate folder, introduced…

introduced test flag

minor bug in printing performances

implemented binary data saving, written test to check multicore performances

start to implement a binary data reader/writer for more memory efficient…

updated gitignore

Pre-Proc working in single/few cores.

still multiproc problems

sync