UTAR Institutional Repository

A framework to automate time series data collection using mobile phone for deep learning training and deployment

Lim, Jason Jing Wei (2019) A framework to automate time series data collection using mobile phone for deep learning training and deployment. Final Year Project, UTAR.

[img]
Preview
PDF
Download (4Mb) | Preview

    Abstract

    Followed by area of machine learning and deep learning, gesture or pattern recognition then became one of the most active subject matter for deep learning. The problem is that data collection process in gesture recognition research is time consuming when it comes to labelling the data. Thus, a framework which facilitates a time series data collection process through dynamic labelling which is useful for deep learning training and deployment is proposed. The framework can be demonstrated through a simple mobile application with the framework implemented within it. First, the Android mobile application collects the time series data with embedded sensors in smartphone and sends the data off to one of the storage service, S3 which is handled by the AWS cloud. On the EC2 handled by AWS, an instance which acts as a web server with the Flask backend is going to receive a HTTP request from the mobile application with the aid of OKHTTPClient, which is a HTTP client that allows HTTP request transmission. After the web server receives the request, it proceeds to build and train the deep learning model based on the data stored in S3. The deep learning model that is going to be used in the project is LSTM model, which is a variant of RNN model. Finally, the web server is going to return the training output to the cloud storage which can be accessed through the mobile application.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: Q Science > Q Science (General)
    Divisions: Faculty of Information and Communication Technology > Bachelor of Computer Science (Hons)
    Depositing User: ML Main Library
    Date Deposited: 20 Aug 2019 12:23
    Last Modified: 20 Aug 2019 12:23
    URI: http://eprints.utar.edu.my/id/eprint/3483

    Actions (login required)

    View Item