(C23-PR569) Project Plan For Bangkit 2023 Product-Based Capstone - Febi Febriyanti Salina Berutu C151DS
(C23-PR569) Project Plan For Bangkit 2023 Product-Based Capstone - Febi Febriyanti Salina Berutu C151DS
(C23-PR569) Project Plan For Bangkit 2023 Product-Based Capstone - Febi Febriyanti Salina Berutu C151DS
Executive Summary/Abstract:
Indonesia is the country with the largest Muslim population in the world, however,
according to data from the Jakarta Institute of Qur'anic Sciences (IIQ), 72% of the Muslim
population cannot read the Al-Qur'an. This phenomenon is certainly very worrying,
moreover it is an important aspect of Islamic worship.
The importance of the Al-Qur'an for Muslims is Firstly, it provides a convenient tool for
Muslims to learn and practice recitation, which is an essential aspect of Islamic worship.
Secondly, the app offers an opportunity for individuals to improve their recitation skills by
providing feedback on their pronunciation and highlighting areas that require improvement.
So this application can improve the ability to read the Al-Qur’an from all circles, not only
among children, but also at the youth and elderly levels with flexible time. That's what we're
concerned about.
Project Schedule:
Based on your team’s knowledge, what tools/IDE/Library and resources that your
team will use to solve the problem?
● Tensorflow Keras:
○ LSTM
● Google Colab
● Python:
○ librosa
○ pandas
○ numpy
○ os
○ matplotlib
○ zipfile
● TensorFlow Lite
● Git/Github
● VS Code:
○ Express as framework
○ eslint as Style Guide
● Google Cloud Platform
○ Cloud Storage
○ Cloud Run / App Engine
● Postman
● Android Studio
● Android Library:
Project Plan for Bangkit 2023 Product-based Capstone
○ Glide
○ Retrofit
○ Lottie
○ JUnit, Mockito
○ Espresso
○ Android KTX
○ Jetpack:
■ Room
■ ViewModel/LiveData
■ Paging
■ Compose
Based on your knowledge and explorations, what will your team need support for?
● Mentors can integrate ML-MD-CC.
● Experienced mentor in building end-to-end machine learning model with a focus on
speech recognition cases and handling audio-type dataset.
● GCP monthly subscription.
● Audio datasets for hijaiyah letters especially harakat (fathah and dhammah); and
each verse of Holy Al-Qur’an.
Based on your knowledge and explorations, tell us the Machine Learning Part of your
Capstone!
Based on our exploration, we will implement a machine learning model to evaluate user
pronunciation when reciting the Holy Al-Quran using speech recognition techniques. The
model will be constructed using the LSTM from TensorFlow Library. The trained model will
be deployed as a mobile application using TFLite for production.
Based on your knowledge and explorations, tell us the Mobile Development Part of
your capstone?
We will work more on the front-end side, starting from design planning, wireframing,
developing application, and app testing. The development and testing is done fully using
Android Studio by utilizing several libraries such as Jetpack and JUnit. The main feature is to
receive user voice input for voice recognition.
Based on your team’s planning, is there any identifiable potential Risk or Issue related
to your project?
● Machine learning model accuracy fall under 80%.
● Machine learning model took longer time to make a prediction.
● Machine learning model to recognise each verse of Holy Al-Qur’an can not perform
well.
● Poor quality of available dataset.
● Insufficient audio dataset, in particular Surah of Holy Al-Qur’an.
● Possible delayed timeline during working period.
● API may have issues when deployed.