537b452449
changed leaky relu syntax to remove keras bug
2026-03-04 17:59:05 +01:00
3169c29319
mini changes in predict pipeline
2026-03-04 17:01:43 +01:00
13bd76631f
outsourcing of scaler in iforest and deep svdd, removal of paths
2026-03-04 16:51:22 +01:00
6cc38291df
clean up predict pipeline 1
2026-03-04 16:21:06 +01:00
8b6c547387
deletion of vae files
2026-03-04 15:19:15 +01:00
de0084dc09
getting rid of redundant files in dataset creation
2026-03-04 15:09:23 +01:00
af3f9d16b2
minor fixes to new paths / dataset with all columns
2026-03-04 12:25:07 +01:00
3d8c7c6639
added predict service, timer and config file from jetson board
2026-03-04 11:14:40 +01:00
7b7f89cf9d
changed gitignore to allow service, timer and yaml
2026-03-04 11:11:29 +01:00
8f24adbdbd
merge of deployment into main
2026-03-04 11:04:00 +01:00
36bae270a1
notebook to calculate replacement values and minor changes
2026-03-04 10:55:56 +01:00
TimoKurz
4cb06d0497
- added comments to the code
...
- translated all outputs and comments to english
2026-03-02 17:55:50 +01:00
f19dde3f9a
update CNNs
2026-03-01 18:31:42 +01:00
TimoKurz
e69000fbd8
- added earlyFusionTest
...
- changed to group split
2026-02-24 19:51:29 +01:00
7a63c7acd3
CNN test and evaluation
2026-02-24 17:54:23 +01:00
0f11a88ae7
CNN with test-split
2026-02-24 15:48:23 +01:00
984ef89a07
change of Early_Fusion
2026-02-24 13:27:12 +01:00
15190ac52e
CNNs
2026-02-24 11:42:58 +01:00
TimoKurz
42965a4733
- added new files
...
- implemented data base writing of AU and ET
2026-02-19 17:14:31 +01:00
4eab3c9876
adjusted paths (this is the deployment setting)
2026-02-16 20:11:07 +00:00
2b01085a9e
scaler v2
2026-02-16 19:32:32 +01:00
0088cef32a
small changes and lazy import of tensorflow
2026-02-16 18:58:18 +01:00
cf88f88814
uploaded config file
2026-02-16 17:07:49 +00:00
2a014e1e4e
first version
2026-02-16 18:06:07 +01:00
TimoKurz
0e25ba4a3e
- added database functionality
...
- added writing of AUs into database in camera_stream
- added script to measure ActionUnits and Eyetracking data in one file
2026-02-15 16:08:23 +01:00
TimoKurz
832a765575
- added implementation to create 50sec video files out of camera lifestream for every 5 seconds
...
- added action unit recognition to the camera stream
2026-02-07 11:27:40 +01:00
3d86bfe6d0
empty feature creation
2026-01-31 17:51:27 +01:00
9b7bb945bc
new files for comissioning
2026-01-29 20:06:42 +00:00
a9ff3880e2
rearrarngement of files
2026-01-29 18:12:13 +01:00
5a216b22fd
tool functions for sqlite database
2026-01-29 17:04:36 +00:00
0294d4e584
files for testing
2026-01-29 13:48:47 +00:00
4f6c3b7370
created low code script for model deployment
2026-01-27 19:28:12 +01:00
9951d8b4f9
aufraeumen branch
2026-01-27 19:10:24 +01:00
5f2db4d0c9
init commit on deployment, removed EDA files
2026-01-27 18:42:40 +01:00
eee173dc0b
added clean eye df call to combined feature creation
...
bug fix, was missing in combined feature creation script
2026-01-26 17:53:54 +01:00
38e9354c42
fixed blink bug in dataset creation
2026-01-26 13:11:06 +01:00
ee648f9adc
added subset filtering to notebook
2026-01-24 19:15:31 +01:00
29d70ce713
created Notebook for subset distribution analysis
2026-01-24 18:15:26 +01:00
d79fad909c
changed script for combined feature creation
...
min duration for blink is now configurable in __main__
2026-01-22 18:43:08 +01:00
TimoKurz
fc6c593f6b
- added xgboost to classify AU and EyeTracking features
2026-01-11 17:21:07 +01:00
10bc7c568a
complete deep svdd training notebook inclucing person wise normalisation
2026-01-02 21:03:58 +01:00
267df136dd
oc com
2026-01-02 13:14:41 +01:00
f0affaed05
AE von ki cluster
2026-01-02 13:14:41 +01:00
a3119e4fc3
tool function for performance based split; new model training notebooks
2026-01-02 12:02:48 +01:00
TimoKurz
f5796a5cdd
- added description to use py-feat
...
- added script:
-- to split a video into 2min clips
-- to extract AUs from a video
2025-12-24 19:08:39 +01:00
a1bdf470cc
changed read me
2025-12-18 15:29:50 +01:00
7708ac1b29
combined dataset creation pipeline
...
CPFC_both.py: creates parquet files chunkwise that include study level phase subjectID, FACE_AU and Eye tracking columns
combined_feature_creation.py : creates a single dataset by computing the features in parallel for the same time window --> same sample rate required
2025-12-18 15:25:18 +01:00
b8bebc0944
minor fixes in dataset creation
...
changed paths to paulusja ... directory
changed feature extraction for AUs to mean instead of sum
added v1 of merge script of datasets (needs to be adjusted)
2025-12-18 13:04:11 +01:00
TimoKurz
87c5e21daf
- changed MAD Outlier Removal - Function
...
- added regulation to xgboost
2025-12-13 13:45:31 +01:00
15b32a9792
removed cell output
2025-12-10 19:35:20 +01:00
