Abstract
Ultrasound imaging is a cornerstone diagnostic tool in healthcare, renowned for its non-invasive, real-time capabilities, portability, and cost-effectiveness. Its role in telemedicine, particularly in remote and resource-limited settings reliant on constrained network infrastructures like 3G, underscores the necessity for efficient image data transmission. However, ultrasound's high data volume and unique imaging characteristics—including speckle noise, spatial resolution, and contrast resolution—pose substantial transmission challenges in low-bandwidth environments.This thesis does not propose a new compression algorithm. Instead, it delivers an engineering benchmark and decision framework for ultrasound image compression in telemedicine: a standardized, metric-driven comparison of representative existing and less explored methods under realistic 3G/4G network assumptions. The study utilizes Breast Ultrasound Images (BUI) and Polycystic Ovary Syndrome (PCOS) datasets to support this evaluation.
Evaluation metrics employed include Full Width at Half Maximum (FWHM) for spatial resolution, Structural Similarity Index (SSIM) for speckle noise preservation, Contrast-to-Noise Ratio (CNR) for contrast resolution, Compression Ratio (CR) for efficiency assessment, and transmission times under simulated 3G and 4G network conditions. The benchmark surfaces method trade-offs and provides recommended method sets aligned to clinical priorities (diagnostic fidelity vs speed-critical triage) and network tiers (3G vs 4G).
This research highlights the importance of compression optimization in enhancing tele-ultrasound capabilities, particularly in rural healthcare delivery. It contributes to the field by identifying and recommending methods that best maintain diagnostic quality while ensuring efficient transmission, thus improving clinical outcomes and healthcare accessibility in bandwidth-limited telemedicine environments. We use “less explored” to denote techniques under-evaluated in ultrasound telemedicine conditions, not unknown in the broader compression literature. The thesis provides a solid foundation for future research, offering practical guidance for medical professionals and policymakers aiming to implement scalable, high-quality telemedicine solutions.
| Date of Award | 2025 |
|---|---|
| Original language | English |
| Supervisor | Dat TRAN (Supervisor), Girija CHETTY (Supervisor) & Dharmendra Sharma AM PhD (Supervisor) |