DLoG – automatic disc based laplacian of gaussian operator for carotid artery boundary extraction in ultrasound images

Q: What is the clinical evidence from this study?

Population: Carotid artery atherosclerosis (ultrasound image segmentation) (n=971). Intervention: Disc-based Laplacian of Gaussian (DLoG) algorithm vs. Existing segmentation methods (e.g., SLIC, ASM, FCM). Primary outcome: Dice Coefficient (DC) for segmentation accuracy.

Q: What does this research mean for the field?

The Disc-based Laplacian of Gaussian (DLoG) algorithm achieves highly accurate carotid artery segmentation in ultrasound images, outperforming existing methods with a Dice Coefficient of 96.4%. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

M Anitha; Arbind Gupta; Prameetha Pai; Rudramurthy V C

doi:10.1038/s41598-026-43716-0

June 22, 2026Scientific ReportsOpen Access

DLoG – automatic disc based laplacian of gaussian operator for carotid artery boundary extraction in ultrasound images

View Full Paper

Ask AI

Bookmark

Share

Key result

DLoG algorithm outperforms existing methods for ultrasound carotid artery segmentation with ~96% accuracy.

n=971

Figures

Why the study?

Does the DLOG algorithm improve the efficiency of carotid artery boundary extraction in ultrasound images compared to other approaches?

Key Points

To develop a robust and accurate disc-based Laplacian of Gaussian operator that incorporates prior shape information for automatic carotid artery boundary segmentation in noisy ultrasound images.
Designed a Disc-based Laplacian of Gaussian (DLoG) operator combining a circular shape prior with enhanced LoG filters at the diameter boundaries and multi-scale radius scanning (radii 20–160 pixels across 10–20 scales).
Applied Gaussian smoothing (scale parameter 0.15rk to 0.25rk) along with a dual-LoG weighting parameter (0.4 to 1.5) to balance noise suppression and circular boundary localization.
Implemented polar transformation, radial 1D filtering (window size 3–7), Savitzky–Golay smoothing (window size 7–21), and outlier removal to extract smooth, gap-free lumen boundaries.
Achieved an overall segmentation efficiency of about 98% across ultrasound datasets with varying speckle noise levels.
Demonstrated accurate center point localization compared with ground truth manual measurements on transverse ultrasound scans.

Structured PICO

Does the DLOG algorithm improve the efficiency of carotid artery boundary extraction in ultrasound images compared to other approaches?

P

Population

Ultrasound images of the common carotid artery (CCA)

I

Intervention

Disc shape based Laplacian of Gaussian (DLOG) algorithm

C

Comparator

Various other segmentation approaches

O

Outcome

Efficiency of boundary extractionsurrogate

The DLOG algorithm provides highly efficient automatic segmentation of the carotid artery in ultrasound images, potentially improving computer-aided diagnosis of cerebrovascular diseases.

Limitations

Assumes near-circular lumen geometry, which may reduce accuracy in cases of severe stenosis or irregular arterial shapes.
Performance can be affected by extremely noisy or low-quality ultrasound images with strong shadowing artifacts.
Not yet optimized or benchmarked on clinical or embedded hardware platforms.

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

Supports automated carotid segmentation research; leaves open clinical validation before diagnostic use.

Abstract

Automatic segmentation of the carotid artery from ultrasound images will provide significant benefit over manual segmentation for assessment and diagnosis of atherosclerosis of the common carotid artery (CCA). The grainy nature of ultrasound images and speckle noise poses several challenges in automatic segmentation of the CCA. To achieve a robust, accurate and reliable segmentation of CCA, a Disc shape based Laplacian of Gaussian (DLOG) is proposed. The proposed approach introduces a disc shape kernel of the size of artery and adding enhanced version of LOG at each end of diameter end of the kernel filter. This priori shape information kernel helps localizing the artery location accurately that looks for global shape features based on intensity changes rather than edge pixels. Gaussian energy term in DLOG operator helps in smoothening that makes the technique more robust to noise and increasing the DLOG efficiency. Further preprocessing operation such as outliner edge removal, boundary gap linking and smoothening is performed to extract the boundary efficiently. The method has been tested on large number of ultrasound images with varying level of noise, to establish its robustness. Experimental results showed that the proposed algorithm has higher efficiency of about 98% compared to various other approaches. Thus, the proposed method shows better performance and has great potential in improving the performance computer-aided diagnosis systems in cerebrovascular diseases.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Anitha et al. (2026) studied Carotid artery atherosclerosis (ultrasound image segmentation) (n=971). Disc-based Laplacian of Gaussian (DLoG) algorithm vs. Existing segmentation methods (e.g., SLIC, ASM, FCM) was evaluated on Dice Coefficient (DC) for segmentation accuracy. The proposed Disc-based Laplacian of Gaussian (DLoG) algorithm achieved a Dice Coefficient of 96.4% for carotid artery segmentation in ultrasound images, outperforming existing methods.

synapsesocial.com/papers/6a3c232cd15afadd906f9d88 https://doi.org/10.1038/s41598-026-43716-0

View Full Paper

Ask AI

Bookmark

Share

DLoG – automatic disc based laplacian of gaussian operator for carotid artery boundary extraction in ultrasound images

Figures

Key Points

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

DLoG – automatic disc based laplacian of gaussian operator for carotid artery boundary extraction in ultrasound images

Figures

Key Points

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading