RbQE: An Efficient Method for Content-Based Medical Image Retrieval Based on Query Expansion.

Area of Science:

• Medical Imaging
• Computer Vision
• Artificial Intelligence

Background:

• The increasing volume of medical images necessitates efficient retrieval systems beyond diagnostics.
• Content-based medical image retrieval (CBMIR) is crucial for understanding disease causes and treatments.

Purpose of the Study:

• To propose an efficient and accurate CBMIR method for computed tomography (CT) and magnetic resonance (MR) images.
• To enhance the understanding of diseases and improve diagnostic capabilities through advanced image retrieval.

Main Methods:

• The proposed RbQE method expands query features using pre-trained AlexNet and VGG-19 models for deep feature extraction.
• RbQE employs a two-stage search: a rapid search for query reformulation and a final search for database retrieval.

Main Results:

• The RbQE method demonstrated superior performance across four public databases (TCIA-CT, EXACT09-CT, NEMA-CT, OASIS-MRI).
• Average retrieval precision (ARP) improvements ranged from 0.84% to 14.34% compared to state-of-the-art methods.

Conclusions:

• The RbQE method offers an efficient and accurate solution for CBMIR of CT and MR images.
• This advancement supports deeper insights into diseases and enhances medical image analysis.