2016
DOI: 10.1016/j.compeleceng.2015.09.013
|View full text |Cite
|
Sign up to set email alerts
|

An efficient two-stage region merging method for interactive image segmentation

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
8
0
1

Year Published

2017
2017
2020
2020

Publication Types

Select...
5
2

Relationship

0
7

Authors

Journals

citations
Cited by 17 publications
(9 citation statements)
references
References 25 publications
0
8
0
1
Order By: Relevance
“…As shown in Figure 1, the flow diagram of the ADRM includes three major components: (1) initial segmentation where over-segmentation is allowed; (2) histogram-based spectral and spatial feature extraction and adaptive region descriptors, and (3) dynamic region merging. The first component can be carried out using some well-known segmentation algorithms, such as mean shift [8,23], watershed [16,24], level set [25], and super-pixel [8]. In this paper, initial segmentation is obtained by using the mean shift method [23].…”
Section: Methodsmentioning
confidence: 99%
See 3 more Smart Citations
“…As shown in Figure 1, the flow diagram of the ADRM includes three major components: (1) initial segmentation where over-segmentation is allowed; (2) histogram-based spectral and spatial feature extraction and adaptive region descriptors, and (3) dynamic region merging. The first component can be carried out using some well-known segmentation algorithms, such as mean shift [8,23], watershed [16,24], level set [25], and super-pixel [8]. In this paper, initial segmentation is obtained by using the mean shift method [23].…”
Section: Methodsmentioning
confidence: 99%
“…By many experiments, the shape parameter is set as 0.1 and the compactness parameter is set as 0.5. As for JSGE, only the parameter "Nu" is chosen by experiments and other parameters are adopted as recommended in the original reference [6] since these parameters can obtain reliable results in various environments [7,8] as well as in our experiments. For SRM, to give a hierarchy of segmentations at different scales, a set of scale parameters (Q − level) are tuned according to the original reference [38], only the one which achieves the best segmentation is selected in this paper as listed in Table 2.…”
Section: Evaluation Methods and Metricsmentioning
confidence: 99%
See 2 more Smart Citations
“…• Stroke(s) [5][6][7][8]: the user is required to place stroke(s) on the object of interest and background on the image • Bounding box [9][10][11]: the user is required to put the bounding box on the object of interest in the image. • Seed point [12,13]: the user is required to put the seed points on the background and object of interest in the image.…”
Section: Introductionmentioning
confidence: 99%