HKU Data Repository
Browse

Supporting data for “Deep learning methods and applications to digital health”

dataset
posted on 2024-10-03, 02:01 authored by Shichao MaShichao Ma

This repository contains three folders which contain either the data or the source code for the three main chapters (Chapter 3, 4, and 5) in the thesis. Those folders are 1) Dataset (Chapter 3): This file contains phonocardigrams signals (/PhysioNet2016) used in Chapter 3 and 4 as the upstream pretraining data. This is a public dataset. /SourceCode includes all the statistical analysis and visualization scripts for Chapter 3. Yaseen_dataset and PASCAL contain phonocardigrams signals with pathological features, Yaseen_dataset serves as the downstream finetuning dataset in Chapter 3, while PASCAL datasets serves as the secondary testing dataset in Chapter 3. 2) Dataset (Chapter 4): /SourceCode includes all the statistical analysis and visualization scripts for Chapter 4. 3) Dataset (Chapter 5): PAD-UFES-20_processed contains dermatology images processed from the PAD-UFES-20 dataset, which is a public dataset. The dataset is used in the Chapter 5. And /SourceCode includes all the statistical analysis and visualization scripts for Chapter 5.

Several packges are mendatory to run the source code, including:

Python > 3.6 (3.11 preferred), TensorFlow > 2.16, Keras > 3.3, NumPy > 1.26, Pandas > 2.2, SciPy > 1.13


History

Usage metrics

    Research Postgraduates

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC