Generating Binary Tags for Fast Medical Image Retrieval Based on Convolutional Nets and Radon Transform

TL;DR

This paper presents a novel medical image retrieval method combining CNN-derived neural codes with Radon barcodes, improving accuracy in large-scale x-ray image archives, and also explores ROI-based search enhancements.

Contribution

The work introduces a new hybrid retrieval approach using CNN and Radon barcodes, demonstrating superior performance on the IRMA dataset.

Findings

Outperforms many published methods on IRMA dataset

Combines global classification with region-based matching

Effective for large medical image archives

Abstract

Content-based image retrieval (CBIR) in large medical image archives is a challenging and necessary task. Generally, different feature extraction methods are used to assign expressive and invariant features to each image such that the search for similar images comes down to feature classification and/or matching. The present work introduces a new image retrieval method for medical applications that employs a convolutional neural network (CNN) with recently introduced Radon barcodes. We combine neural codes for global classification with Radon barcodes for the final retrieval. We also examine image search based on regions of interest (ROI) matching after image retrieval. The IRMA dataset with more than 14,000 x-rays images is used to evaluate the performance of our method. Experimental results show that our approach is superior to many published works.