This paper proposes an active contour driven by adaptively weighted signed pressure force (SPF) combined with the Legendre polynomial method for image segmentation. First, an adaptively weighted global average intensity (GAI) term is defined wherein GAI differences are the weighted factors of the interior and exterior region-driving centers. Second, an adaptively weighted Legendre polynomial intensity (LPI) term is defined which adopts the Legendre polynomial intensity average differences as the weighted factors of the interior and exterior region-driving centers. Finally, the GAI and LPI terms are introduced into a novel SPF function and a coefficient is applied to weight their effect degrees; a new edge stopping function (ESF) is defined and combined with the region-based method to robustly converge the curve to the boundary of the object. Experiments demonstrate that this method is highly accurate and computationally efficient for images with inhomogeneous intensity, blurred edge, low contrast, and noise problems. Moreover, the segmentation results are independent of the initial contour.
Bibliographical noteThis research is supported by the National Natural Science Foundation of China (61876026, 61906022, 61703062, 61871342, 61772344, 61672443), the Chongqing Special Key Project of Technology Innovation and Application Development (cstc2019jscx-mbdxX0027), the Fundamental Research Funds for the Central Universities (2020CDCGJSJ039, 2020CDJ-LHZZ-052), in part by the General Program of National Natural Science Foundation of Chongqing (cstc2020jcyj-msxmX0790), in part by the Guangxi Key Laboratory of Cryptography and Information Security (GCIS201906), in part by the Hong Kong RGC General Research Funds (9042820 (CityU 11219019), 9042489 (CityU 11206317), 9042322 (CityU 11200116), 9048123 (CityU 21211518), in part by the Key Project of Science and Technology Innovation 2030 supported by the Ministry of Science and Technology of China (2018AAA0101301, i).
- Active contour
- Image segmentation
- Level set
- SPF function