Otsu's Objective Research Articles

Multilevel thresholding selection based on variational mode decomposition for image segmentation

Multilevel thresholding techniques based on gray histogram are usually computationally expensive for the image segmentation. In this paper, we propose a novel thresholding extraction method based on variational mode decomposition (VMD). The improved VMD is used to decompose the histogram non-recursively into several sub-modes for minimizing Otsu's objective function. Then, we can extract the thresholds easily by the minimum point search (MPS) method or the cross point search (CPS) method. The experimental results demonstrate that the proposed MPS scheme exhibits more excellent capability than CPS. Further, compared with other approaches namely particle swarm optimization algorithm (PSO), fuzzy c-means (FCM) algorithm and bacterial foraging (BF) algorithm, the proposed algorithm can get similar performance, but its computing speed is faster than others. Therefore, it could have some advantages in image preprocessing, such as fast target recognition and classification.

Multi-objective and multi-level image thresholding based on dominance and diversity criteria

The Kapur and Otsu methods are widely used image thresholding approaches and they are very efficient in bi-level thresholding applications. Evolutionary algorithms have been developed to extend the Kapur and Otsu methods to the multi-level thresholding case. However, there remains an unsolved argument that neither Kapur nor Otsu objective can optimally fit diverse content contained in different kinds of images. This paper proposes a multi-objective model which seeks to find the Pareto-optimal set with respect to Kapur and Otsu objectives. Based on dominance and diversity criteria, we developed a hybrid multi-objective particle swarm optimization (MOPSO) method by incorporating several intelligent search strategies. The ensemble strategy is also applied to automatically select the best search strategy to perform at various algorithm stages according to its historic performances. The experimental result shows that the solutions to our multi-objective model consistently produce equal or better segmentation results than those by the optimal solutions to the original Kapur and Otsu models, and that the proposed hybrid algorithm with and without the ensemble strategy produces a better approximation to the ideal Pareto front than those obtained by two other MOPSO variants and the MOEA/D. In comparison with the most recent multilevel thresholding methods, our approach also consistently obtains better performance in the segmentation result for several benchmark images.

A hybrid flower pollination algorithm based modified randomized location for multi-threshold medical image segmentation.

Multi-threshold image segmentation is a powerful image processing technique that is used for the preprocessing of pattern recognition and computer vision. However, traditional multilevel thresholding methods are computationally expensive because they involve exhaustively searching the optimal thresholds to optimize the objective functions. To overcome this drawback, this paper proposes a flower pollination algorithm with a randomized location modification. The proposed algorithm is used to find optimal threshold values for maximizing Otsu's objective functions with regard to eight medical grayscale images. When benchmarked against other state-of-the-art evolutionary algorithms, the new algorithm proves itself to be robust and effective through numerical experimental results including Otsu's objective values and standard deviations.

Modified particle swarm optimization-based multilevel thresholding for image segmentation

Since the conventional multilevel thresholding approaches exhaustively search the optimal thresholds to optimize objective functions, they are computational expensive. In this paper, the modified particle swarm optimization (MPSO) algorithm is proposed to overcome this drawback. The MPSO employs two new strategies to improve the performance of original particle swarm optimization (PSO), which are named adaptive inertia (AI) and adaptive population (AP), respectively. With the help of AI strategy, inertia weight is variable with the searching state, which helps MPSO to increase search efficiency and convergence speed. Moreover, with the help of AP strategy, the population size of MPSO is also variable with the searching state, which mainly helps the algorithm to jump out of local optima. Here, the searching state is estimated as exploration or exploitation simply according to whether the gBest has been updated in $$k$$ k consecutive generations or not, where the gBest stands for the position with the best fitness found so far among all the particles in the swarm. The MPSO has been evaluated on 12 unimodal and multimodal Benchmark functions, and the effects of AI and AP strategies are studied. The results show that MPSO improves the performance of the PSO paradigm. The MPSO is also used to find the optimal thresholds by maximizing the Otsu's objective function, and its performance has been validated on 16 standard test images. The experimental results of 30 independent runs illustrate the better solution quality of MPSO when compared with the global particle swarm optimization and standard genetic algorithm.