adamjeddy/grad-project-code

Repo for storing files for the graduation project. It holds code for the CV and NLP Model

Activity Metrics	Language & Repository Metrics

grad-project-code's Issues

Dataset - check if there is thank you; replace with "Thanks" because that is the training word for CV model

Change the decoder function code to take user inputs

NLP_Model.py file should be able to take an ASL input and it should give out an output

NLP Model can't handle words that aren't part of the dataset

Fix will most likely take place in the decoder function.

Test NLP Model

Try out different inputs for the model

It tends to break when a word that's not part of the dataset it entered

Incorrect preprocessing of Dataset

The new dataset needs to be preprocessed again because a row has more than 1 comma.
This causes issues since it will be used as a CSV (comma-separated values) file.

Script: removing the script which handles the 6W's and H

because in gloss they normally use the 6W's, even if they are not asking a question.
e.g.
My favorite class is science, MY FAVORITE CLASS WHAT SCIENCE
Today the weather is sunny, WEATHER TODAY WHAT SUNNY

Getting finger words into 1 dictionary

Creating a conversational dataset for training of CV and NLP model

Fix Paths/Links of all files

The repository was reorganized so the links/paths used in the notebook (and scripts) are invalid/broken

Script: Replace words from dictionary with a place holder

NLP: Handling '?' (Question mark in ASL)

Change all Hi/Hey to Hello in the NLP Dataset

Check Dataset for no spaces between words

Examples:
your resume howmany page
nextyear school register finish you
you like takecareof plant you

Check on this

Add finger spelling handling in script

Have a placeholder for nouns and then change it

Input ASL: I love Bob -> I love X
Output English: I love X -> I love Bob

Testing: How are finger sign is going to be outputted