Intravascular ultrasound image recovery and segmentation based on circular analysis

Abstract

This work presents a novel approach to recover incomplete intravascular ultrasound image before segmenting the lumen and media-adventitia boundaries. The proposed algorithm used circular analysis to estimate lumen and media-adventitia boundaries and then applied diffusion technique to recover corrupted pixel within the detected area. The proposed method was tested with the standard dataset provided by the computer vision center, Universitat de Barcelona, Barcelona, ESP. The experimental result demonstrates that the proposed method provided an average improvement rate of 11% compared to others method on Jaccard index, 9% on Hausdorff distance, 12% on Percentage Area Distance, 6.2% on Area Overlap Error.