.png){kind=icon}
.png){kind=icon}
Return
Parametric active contour based on sparse decomposition for multi-objects extraction
.png){kind=icon}
PRE
AI
DOI：10.1016/j.sigpro.2018.02.032.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Active contour model has been widely used over the past decade in image segmentation. For parametric active contour model, it is always used to segment objects because of its simplicity and fast evolution. However, it could not extract multi-objects with one initial contour because of topological invariance of parametric contour and existing equilibrium points between objects in vector fields. In this paper, a method of multi-objects extraction with parametric active contour model is proposed. Firstly, after an edge map is computed, the idea of sparse representation and decomposition for edge map is introduced in order to obtain new edge maps to better describe the objects; Secondly, the obtained edge maps are used to generate vector fields; Finally, one initial contour is evolved in every vector field to extract corresponding objects. Experimental results show that the proposed model could extract multi-target objects by using one initial contour. Quantitative evaluation of the segmentation results with tested methods also shows that the proposed method is more robust to noise and gets the better effect of segmentation accuracy. Furthermore, the proposed method allows parallel computation of algorithms, hence further reducing the computational time. (C) 2018 Elsevier B.V. All rights reserved.
Keywords:
Parametric active contour model
Multi-objects
Vector field
Edge map
Sparse decomposition
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.