Comparative evaluation of fractal analysis and canny edge-detection algorithms for the quantification and classification of bone trabecular patterns in cone-beam computed tomography
4 affiliations
- 1Department of Oral and Maxillofacial Radiology, Dental Sciences Research Center, School of Dentistry, Guilan University of Medical Sciences
- 2Department of Periodontics, Dental Sciences Research Center, School of Dentistry, Guilan University of Medical Sciences
- 3Department of Prosthodontics, Dental Sciences Research Center, School of Dentistry, Guilan University of Medical Sciences
- 4Department of Biostatistics and Epidemiology, School of Health, Qom University of Medical Sciences
Abstract
While gray scale density metrics are susceptible to cone beam computed tomography (CBCT) noise and calibration inconsistencies, advanced image processing algorithms offer a potential non-invasive, objective alternative. This study presents a comparative evaluation of Fractal Analysis (FA) and Canny Edge-Detection (CED) algorithms for quantifying and classifying distinct bone trabecular patterns. In this ex vivo experimental study, 120 bovine rib bone blocks were categorized into four morphologic groups (D1-D4) and imaged using a standardized CBCT protocol. A single standardized 50 × 50-pixel region of interest (ROI) was analyzed on a selected two-dimensional reconstructed slice from each block. FA and CED quantified structural complexity and edge count, respectively. Group differences were evaluated using analysis of variance with multiplicity-adjusted post hoc tests, and exploratory receiver operating characteristic (ROC) analyses were used to estimate pairwise discrimination. Both FA and CED values differed significantly across the four morphologic groups (p < 0.001). FA differentiated every pairwise comparison (p ≤ 0.014), whereas CED did not differentiate D2 from D3 (p = 0.717). Exploratory AUCs ranged from 0.77 to 0.99 for FA and from 0.64 to 1.00 for CED. Within this controlled ex vivo model, FA and CED provided objective, reproducible quantification of distinct trabecular bone patterns on CBCT images. FA excels in differentiating intermediate densities (D2 and D3), while CED offers superior performance for denser and looser bone patterns (D1 and D4). The complementary performance of the two algorithms supports their potential as quantitative adjuncts to visual bone-pattern assessment.
Related
- Comparative Evaluation of Feature Selection Algorithms for Predicting Radiation-Induced Complications in Head and Neck Cancer Using Multimodal CT and MRI Radiomics
- Comparative Evaluation of Multiple Deep Learning Models for Six-Class Keratitis Classification
- Deep Learning-Based Classification of Melanoma Using Dermoscopic Images: Comparative Evaluation of Convolutional Neural Network and DenseNet Models.
- Comparative Evaluation of EfficientNet-B3, Vision Transformers, and RETFound for Seven-Category Retinopathy Fundus Image Classification.
- Evaluation of iodine detection and quantification limits in photon-counting CT—a phantom study
BibTeX
@article{habibi2026comparative,
title = {{Comparative evaluation of fractal analysis and canny edge-detection algorithms for the quantification and classification of bone trabecular patterns in cone-beam computed tomography}},
author = {Kasra Habibi and Bardia Vadiati Saberi and Amirreza Hendi and Negar Khosravifard and Farzaneh Ostovarrad and Mohammad Ebrahim Ghaffari},
journal = {Scientific Reports},
year = {2026},
month = sep,
doi = {10.1038/s41598-026-72893-1},
url = {https://doi.org/10.1038/s41598-026-72893-1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Comparative evaluation of fractal analysis and canny edge-detection algorithms for the quantification and classification of bone trabecular patterns in cone-beam computed tomography
AU - Kasra Habibi
AU - Bardia Vadiati Saberi
AU - Amirreza Hendi
AU - Negar Khosravifard
AU - Farzaneh Ostovarrad
AU - Mohammad Ebrahim Ghaffari
PY - 2026
DA - 2026/09/01
JO - Scientific Reports
DO - 10.1038/s41598-026-72893-1
UR - https://doi.org/10.1038/s41598-026-72893-1
AB - While gray scale density metrics are susceptible to cone beam computed tomography (CBCT) noise and calibration inconsistencies, advanced image processing algorithms offer a potential non-invasive, objective alternative. This study presents a comparative evaluation of Fractal Analysis (FA) and Canny Edge-Detection (CED) algorithms for quantifying and classifying distinct bone trabecular patterns. In this ex vivo experimental study, 120 bovine rib bone blocks were categorized into four morphologic groups (D1-D4) and imaged using a standardized CBCT protocol. A single standardized 50 × 50-pixel region of interest (ROI) was analyzed on a selected two-dimensional reconstructed slice from each block. FA and CED quantified structural complexity and edge count, respectively. Group differences were evaluated using analysis of variance with multiplicity-adjusted post hoc tests, and exploratory receiver operating characteristic (ROC) analyses were used to estimate pairwise discrimination. Both FA and CED values differed significantly across the four morphologic groups (p < 0.001). FA differentiated every pairwise comparison (p ≤ 0.014), whereas CED did not differentiate D2 from D3 (p = 0.717). Exploratory AUCs ranged from 0.77 to 0.99 for FA and from 0.64 to 1.00 for CED. Within this controlled ex vivo model, FA and CED provided objective, reproducible quantification of distinct trabecular bone patterns on CBCT images. FA excels in differentiating intermediate densities (D2 and D3), while CED offers superior performance for denser and looser bone patterns (D1 and D4). The complementary performance of the two algorithms supports their potential as quantitative adjuncts to visual bone-pattern assessment.
SN - 2045-2322
ER -