Effective image clustering based on human mental search

Abstract

Image segmentation is one of the fundamental techniques in image analysis. One group of segmentation techniques is based on clustering principles, where association of image pixels is based on a similarity criterion. Conventional clustering algorithms, such as k-means, can be used for this purpose but have several drawbacks including dependence on initialisation conditions and a higher likelihood of converging to local rather than global optima. In this paper, we propose a clustering-based image segmentation method that is based on the human mental search (HMS) algorithm. HMS is a recent metaheuristic algorithm based on the manner of searching in the space of online auctions. In HMS, each candidate solution is called a bid, and the algorithm comprises three major stages: mental search, which explores the vicinity of a solution using Levy flight to find better solutions; grouping which places a set of candidate solutions into a group using a clustering algorithm; and moving bids toward promising solution areas. In our image clustering application, bids encode the cluster centres and we evaluate three different objective functions. In an extensive set of experiments, we compare the efficacy of our proposed approach with several state-of-the-art metaheuristic algorithms including a genetic algorithm, differential evolution, particle swarm optimisation, artificial bee colony algorithm, and harmony search. We assess the techniques based on a variety of metrics including the objective functions, a cluster validty index, as well as unsupervised and supervised image segmentation criteria. Moreover, we perform some tests in higher dimensions, and conduct a statistical analysis to compare our proposed method to its competitors. The obtained results clearly show that the proposed algorithm represents a highly effective approach to image clustering that outperforms other state-of-the-art techniques.

AMC2-Pyramid: Intelligent Pyramidal Feature Engineering and Multi-Distance Decision Making for Automatic Multi-Carrier Modulation Classification

“…In the proposed TF-HMS, a form of a bid is feature vectors , which encode the cluster centre. An array length is represented as , which is a number of clusters [48]. The upper and lower bound values are computed for each bid, which is represented as = min ( ), = max ( ), where the upper and lower bounds are the minimum-and maximum-based feature vectors.…”

“…The weight value is computed by all distance values as follows, which is the sum of all distance measures. = 1 ( , )+ 2 ( , )+ 3 ( , )+ 4 ( , ) 4 0 ≤ ≤ 1 (48) where 1 , 2 , 3 and 4 are the distance values of , , and between and clusters, respectively. Based on , the nearest centroid is computed for the particular signal.…”

Automatic clustering using a local search-based human mental search algorithm for image segmentation

“…Clustering is applied to identify similar groups (clusters) in an image so that the samples associated with a group share similar features, while the samples belonging to different groups have different features. Clustering in images is challenging in comparison to clustering other types of data [6] because the number of samples (pixels) in an image is high since even a relatively small image of size 256 × 256 has 65,536 pixels, while an 512 × 512 image has 262,144 samples.…”

“…The algorithm has three main stages: mental search, which searches around candidate solutions using a Levy flight mechanism; bid grouping, which finds promising areas using a clustering algorithm; and moving candidate solutions towards promising areas. HMS has been shown to provide competitive performance compared to other metaheuristic algorithms [6,35,36] due to its powerful operators. For example, the update in HMS algorithm is based on a region rather than a point which better allows HMS to prevent early convergence.…”

“…In our earlier work, we have applied HMS to image clustering in [6] which has demonstrated the algorithm's ability to improve upon other image clustering approaches. There, our encoding representation was a one-dimensional array defining the cluster centres, while three different cost functions were evaluated.…”