mobile-vsr

This is the repository for mobile based lipreading, "MobiVSR v1".

Demo Link and Instructions

Get the demo for MacOS at the following Link: https://drive.google.com/file/d/1MTgX9fnZSvTng4S5FeVDJB-Y12cLnZkm/view?usp=sharing
Run the executable.
Press Detect face button to automatically detect face. (try multiple times for optimal results)
Press the start button and speak any words from the list given in the file "label_sorted.txt"
The recorder automatically closes after 1.19 seconds and displays a list of top 5 predictions

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
docs		docs
README.md		README.md
auxillary-files-mobivsr.zip		auxillary-files-mobivsr.zip
config.json		config.json
cvtransforms.py		cvtransforms.py
dataset.py		dataset.py
depthwise.py		depthwise.py
label_sorted.txt		label_sorted.txt
lr_scheduler.py		lr_scheduler.py
main.py		main.py
model.py		model.py
utils.py		utils.py