Fuzzy C-Means Image Segmentation Algorithm Based on Chaotic Simulated Annealing

Abstract

Considering the problem that the traditional fuzzy c-means (FCM) image segmentation algorithm is often caught in a specific range in local search and fails to get the globally optimal solution, this paper proposed a modified FCM algorithm based on chaotic simulated annealing (CSA). It traverse all the states without repetition within a certain range to calculate the optimal solution. Experimental results show that our method converges more quickly and accurately to the global optimal and proves a promise global optimization method of high adaptability and feasibility.

Development and implementation of (Q)SAR modeling within the CHARMMing web-user interface

“…[19] We have previously developed SAR models using two advanced ML classifiers: random forest (RF) [20,21] and k Nearest Neighbor Simulated Annealing. [22] We identified novel nonnucleoside chemical motifs and Candesartan cilexetil (a drug used to treat hypertension and heart failure) [23] for NS5B, hepatitis C virus RNA polymerase activity. Herein, we describe the development of a general (Q)SAR infrastructure based around the CHARMM web-user interface, charmming.org to be used to build these types of models.…”

“…[27] ChemBench offers a convenient user interface and a way to set up cross validation, as well as track submitted jobs. However, it has only 4 ML methods (RF, [20,21] support vector machine, [28] kNN-SA, [22] and kNN-GA [29] ). It also seems to be quite slow (a submitted job can take hours and even days to complete) and requires some data preparation before use-the activities and structures have to be uploaded as separate files.…”