BabyReady uses a Convolutional Neural Network to predict the reason a baby is crying with a mp3 recording as input.
BabyReady was developed in 2018 by Ang Kang Rong Roy
and Kyle Zheng Ching Chan
with help from Tang Yetong
. It won the champion and healthtech award at Splash Awards 2018 and first runner-up in CodeXtremeApps 2018.
- Run the BabyReady android app and place device beside baby.
- When baby cries, the cry is recorded and uploaded to the online server which predicts the reason
- The reason is sent to the android device through an online database and the device will act based on the predicted reason. If the baby is scared or lonely, Baby Shark is played. If the baby is tired, a lullaby is played.
You must install both the server and android app.
- Download the entire
BabyReady
repository. - Create your own Firebase project and download the
google-servies.json
file. Add it to bothBabyReady/BabyReady - Python/auth.json
andBabyReady/BabyReady - Android/app/google-services.json
. Create an environmental variable on your computer GOOGLE_APPLICATION_CREDENTIALS and enter as value the file location ofBabyReady/BabyReady - Python/auth.json
. pip install -r requirements.txt
.- Build the apk from the Android project and install in android device.
- Add data to
BabyReady/BabyReady - Python/data/
- Run
BabyReady/BabyReady - Python/train_set.py
- Run
BabyReady/BabyReady - Python/train_model.py
A sample model has already been created in BabyReady/BabyReady - Python/output/
. You may continue with this if you do not want to create a new model yourself.
Start server by running BabyReady/BabyReady - Python/try_firebase.py
on your computer
Start client by running android apk
built from Android Project in BabyReady/BabyReady - Android/
Alternatively, just run BabyReady/BabyReady - Python/BabyCryRecording.py on your computer
Codes for each reason for baby cry: bp - belly pain
bu - burp
ch - temperature
dc - discomfort
hu - hungry
lo - lonely
sc - scared
ti - tired