Neutrosophic Set Transformation Matrix Factorization Based Active Contours for Color Texture Segmentation

Abstract

Active contour model (ACM) has widely used for segmenting two-phase images. However, its performance may not be satisfactory for some color texture images when their features cannot be effectively extracted. To alleviate this problem, in this paper, a novel neutrosophic set transformation matrix factorization-based active contour (NSTMF-AC) approach is proposed for color texture segmentation. The proposed NSTMF-AC is an effective and robust color texture segmentation method. Particularly, to effectively capture a wide range of texture information, the proposed method extracts the features from the triple domains, including spatial, wavelet, and spectral domains, and then uses neutrosophic set (NS) transform and the corresponding operations to reduce the indeterminacy contained in the image. Furthermore, the method obtains the resulting NS transformation matrix and utilizes a factorization-based ACM to perform image segmentation. Finally, the proposed method is compared with the state-of-the-art segmentation algorithms on a variety of natural images. The experimental results demonstrate that the proposed NSTMF-AC method is more robust for two-phase image segmentation than other methods. INDEX TERMS Texture segmentation, neutrosophic set, matrix factorization, active contour model.

Toward an Optimized Neutrosophic k-Means With Genetic Algorithm for Automatic Vehicle License Plate Recognition (ONKM-AVLPR)

“…So, for the image, the pixel A(x, y) is defined as A NS (x, y) = A (t, i, f) = {T (x, y), I(x, y), F(x, y)} for (NS) domain giving the true, indeterminate, and false belonging to the bright pixel set. Assume A(x, y) demonstrate the intensity value of the pixel (x, y), and A(x, y) indicated to its local mean value, the membership functions can be represented as follows [21]- [24].…”

GPU-accelerated image segmentation based on level sets and multiple texture features

“…Gao et al introduced a factorisation-based ACM that utilises the local spectral histogram as the texture feature [16], and, more recently, a model that performs a fusion of intensity and Gabor-based features along with a factorisation scheme [17]. Dong et al [14] also employed a factorisation-based ACM, combined with neutrosophic sets, in the task of color texture segmentation. Dahl and Dahl [11] created a method based on ACWE and probabilistic image patch dictionaries.…”

“…• 100 (14) This term gives a different response in each of the image points. Its main idea is to generally discourage the contour from expanding to the areas with a large number of sufficiently different features.…”

“…Texture segmentation using active contour models has been a popular topic of research in recent years [1,14,17,51,52]. Particularly, the variational ACWE-like models and their modified multi-channel versions for vector-valued images [8] have been often employed [1,19,32,57].…”

Multi-Resolution Texture-Based 3D Level Set Segmentation

“…For regions without clear borders or with non-uniform texture, however, these approaches are usually inadequate. To solve the problem, many texture feature extraction techniques, such as grey level co-occurrence matrices (GLCM) [8], [9], Gabor filters [10], matrix factorisation [11], [12], or wavelets [13], are integrated with deformable models. Typically, these methods use a single feature [11], [14], [15] or ad hoc combinations of descriptors [16], [17].…”