Image separation is a significant task concerned in dissimilar areas from image dispensation to picture examination. One of the simplest methods for image segmentation is thresholding. Though, many thresholding methods are based on a bi-level thresholding process. These methods can be extensive to form multi-level thresholding, but they become computationally expensive since a large number of iterations would be necessary for computing the most select threshold values. In order to conquer this difficulty, a new process based on a Shrinking Search Space (3S) algorithm is proposed in this paper. The method is applied on statistical bi-level thresholding approaches including Entropy, Cross-entropy, Covariance, and Divergent Based Thresholding (DBT), to attain multi-level thresholding and used for separation from brain MRI images. The paper demonstrates that the collision of the proposed 3S method on the DBT method is more important than the other bi-level thresholding approaches. Comparing the results of using the proposed approach against those of the Fuzzy C-Means (FCM) clustering method demonstrates a better segmentation performance by improving the comparison index from 0.58 in FCM to 0.68 in the 3S method. Also, this method has a lower calculation impediment of around 0.37s with admiration to 157s dispensation time in FCM. In addition, the FCM approach does not always guarantee the convergence, whilst the 3S method always converges to the optimal result.
Adaptive Feature Pyramid Network to Predict Crisp Boundaries via NMS Layer and ODS F-Measure Loss Function
