DMpDP: a Diagnostic Multiple-patient DermoFeature Profile store-and-forward teledermoscopy system

Area of Science:

• Medical Imaging
• Telehealth Technology
• Signal Processing

Background:

• Growing demand for telehealth necessitates efficient, scalable solutions for simultaneous patient care.
• Teledermoscopy requires effective methods for transmitting large image files over networks.
• Existing systems face challenges in bandwidth and data management for multiple patients.

Purpose of the Study:

• To develop a store-and-forward (SAF) teledermoscopy system utilizing compressed dermoFeature profiles (DPs).
• To propose and evaluate different arrangements of Diagnostic Multiple-patient DermoFeature Profiles (DMpDPs) for enhanced telehealth transmission.
• To optimize the embedding of DMpDPs within patient information signals for improved diagnostic accuracy and signal quality.

Main Methods:

• Development of a dermoFeature profile (DP) by extracting weighted intensity-difference frequency (WIDF) and morphological features (MOFs) to reduce image size.
• Assembly of DPs into horizontally aligned, diagonal-based, and sequential-based Diagnostic Multiple-patient DermoFeature Profiles (DMpDPs).
• Embedding DMpDPs into a recorded patient-information signal (RPS) with a weight factor (β) and evaluating performance under various conditions.

Main Results:

• The sequential-based DMpDP arrangement demonstrated superior diagnostic classification accuracy.
• High accuracy was achieved under -5 dB additive white Gaussian noise, with a 98.79% signal-to-noise ratio.
• Optimal performance was observed with β=100, transmitting 248 DPs, utilizing spectral subtraction filtering.

Conclusions:

• The proposed DMpDP approach significantly enhances the efficiency of SAF teledermoscopy systems.
• The sequential-based DMpDP is effective for improving diagnostic accuracy in noisy telehealth environments.
• This method offers a scalable solution for transmitting multiple patient dermoscopy images in real-time healthcare applications.