Related Experiment Video
Updated: Aug 10, 2026
05:24
Multifractal Spectrum Analysis for Assessing Pulmonary Nodule Malignancy
Published on: January 10, 2025
Pulmonary nodule detection in CT images with quantized convergence index filter
Sumiaki Matsumoto1, Harold L Kundel, James C Gee
1Department of Radiology, Kobe University Graduate School of Medicine, Kobe, Hyogo 650-0017, Japan. sumatsu@med.kobe-u.ac.jp
Medical Image Analysis
|March 18, 2006
Summary
A new quantized convergence index (QCI) filter improves detection of rounded lesions in CT scans. This computer-aided diagnosis scheme enhances pulmonary nodule identification with high sensitivity and few false positives.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Radiology
Background:
- Pulmonary nodules require accurate detection in CT images for timely diagnosis.
- Existing computer-aided diagnosis (CAD) schemes can be improved for nodule conspicuity.
Purpose of the Study:
- To introduce a novel quantized convergence index (QCI) filter for enhancing rounded lesions.
- To integrate the QCI filter into a CAD scheme for improved pulmonary nodule detection in CT images.
Main Methods:
- Developed a QCI filter, building upon the convergence index (CI) filter, to quantify gradient convergence.
- Evaluated the QCI filter and a CAD scheme using five clinical datasets with 50 nodules.
- The CAD scheme utilized QCI filter output for candidate generation and linear discriminant analysis for classification.
Main Results:
- The QCI filter demonstrated a more selective response to nodules compared to the CI filter using a 9x9 support region.
- The CAD scheme achieved 90% sensitivity with 1.67 false positives per slice.
- The QCI filter output level was the most effective feature for classifying nodule candidates.
Conclusions:
- The QCI filter is a promising tool for preprocessing in automated pulmonary nodule detection.
- The proposed CAD scheme effectively utilizes the QCI filter for enhanced nodule detection in CT images.
