Image decomposition-based sparse extreme pixel-level feature detection model with application to medical images

Lahoti, Geet; Chen, Jialei; Yue, Xiaowei; Yan, Hao; Ranjan, Chitta; Qian, Zhen; Zhang, Chuck; Wang, Ben

doi:10.1080/24725579.2021.1910599

Cited by 2 publications

(1 citation statement)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the smoothness of the background, a spline basis will be taken into account. In terms of the striations area, the prior information from lofargram shows that the striations are wide, which indicates that the spline basis is a preferable choice [15]. In order to estimate the model parameters, θ and

\theta _a

, the least square regression is used.…”

Section: Methodsmentioning

confidence: 99%

A method for extracting interference striations in lofargram based on decomposition and clustering

Wang

Tian

et al. 2023

IET Image Processing

View full text Add to dashboard Cite

In complex underwater noise environment, target detection, recognition, and tracking are proceeded through the frequency spectrum analysis of the signals received by sonar systems, in which the lofargram plays a major role. Typically, if the distance of the target experiences far‐near‐far variation, there will be parabolic interference striations presented in the lofargram. However, the existing striations extraction methods sometimes lack objectivity and fail to extract the wide striations accurately. In this paper, a novel method for wide type interference striations extraction is developed based on efficient decomposition and ensemble clustering. To obtain valuable information, the lofargram is decomposed into smooth background, striations area, and noise, then a multi‐phase ensemble clustering algorithm is employed to extract parabolas from the decomposed striations area pixels. Experimental results on simulated and real‐life datasets exhibit the effectiveness of the proposed method, and verify that it has comparable performance with prevalent Hough transform while extracting striations.

show abstract