search

dattalab / moseq2-app

MoSeq2 Jupyter Notebook platform used to run all of the MoSeq2 tools in a GUI.
Other
32 stars 10 forks source link

Cable filter doesn't work well #245

Open EstherXSD opened 1 year ago

EstherXSD commented 1 year ago

Hi, this is Shudi from the Uchida Lab. We previously did fiber photometry with novelty task and planned to analyzed the behavioral videos. In order to filter the cable, we used the cable-filter-iters 1. However, the cable was still obvious in the processed video.

Expected Behavior

I expect the cable-filter-iters would filter the cable.

Current Behavior

The cable-filter-iters doesn't filter the cable.

for dir in session*/; do moseq2-extract extract $dir/depth.dat --flip-classifier /home/alex/moseq2/flip_classifier_k2_c57_10to13weeks.pkl --bg-roi-dilate 75 75 --use-tracking-model True --cable-filter-iters 1 done

The result video: https://drive.google.com/file/d/1BceRlmJ3SG1W11or2xcshDG5kAbjLDPY/view?usp=sharing The process configuration: https://drive.google.com/file/d/1qJcAszWhBJpMwS9HV5A6EnDJz-FHQLFm/view?usp=sharing

Your Environment

NadineKGut commented 1 year ago

We have the same problem (Mena Lab). Unfortunately, none of the filter sizes nor changes in the parameters of the tracking model could fix that. We just proceeded by discarding the worst videos. But we are still struggling to get the explained variance by first 5-10 syllables above 80%.