Applied Soft Computing 11 (2011) 4883–4893 Contents lists available at ScienceDirect Applied Soft Computing j ourna l ho mepage: www.elsevier.com/locate/asoc An ant-inspired algorithm for detection of image edge features S. Ali Etemad a,∗ , Tony White b a Department of Systems and Computer Engineering, Carleton University, Ottawa, Ontario K1S5B6, Canada b Department of Computer Science, Carleton University, Ottawa, Ontario K1S5B6, Canada a r t i c l e i n f o Article history: Received 21 November 2010 Received in revised form 14 May 2011 Accepted 14 June 2011 Available online 24 June 2011 Keywords: Ant colony systems Feature extraction Image edge analysis a b s t r a c t This paper presents a technique inspired by swarm methodologies such as ant colony algorithms for processing simple and complicated images. It is shown that the proposed technique for image processing is capable of performing feature extraction for edge detection and segmentation, even in the presence of noise. Our proposed approach, Ant-based Correlation for Edge Detection (ACED), is tested on different samples and the results are compared to typical established non-swarm-based methods. The comparative analysis highlights the advantages of the proposed method which generates less distortion when noise is added to the test images. Both qualitative and quantitative evaluations support the claim, confirming the significance of our swarm-based method for image feature extraction and segmentation. © 2011 Elsevier B.V. All rights reserved. 1. Introduction Image processing and machine vision have for long been a vital element in various fields of technology. While automated and artificially intelligent systems in many cases require a signif- icant amount of precise and robust image processing capabilities, robust techniques are yet to be proposed and demonstrated for many ongoing problems. Different techniques ranging from sim- ple filtering procedures to more complicated methods like SIFT [1] aid in solving many practical problems for industrial and research applications. In this regard, more dynamic algorithms, despite their promising performance in many other fields, have not been widely explored. This could be in some part due to the existing gap between researchers in the two fields of machine vision and soft computing. Swarm intelligence is a branch of artificial intelligence which provides dynamic, adaptive, and flexible solutions using self- organized decentralized algorithms [2]. While swarm intelligence maintains stochastic properties for most of its procedures, it man- ages to provide robust solutions in many domains [3,4,5]. The applications of swarm intelligence for practical research and even industrial problems have grown substantially in recent years. In this paper, we approach the problem of image feature extrac- tion using swarm-based image processing. We propose a method capable of extracting image features for edge detection. The method is employed and tested on different images, as well as the bench- ∗ Corresponding author. E-mail addresses: etemad@sce.carleton.ca (S.A. Etemad), arpwhite@scs.carleton.ca (T. White). mark test on Lena. Furthermore, sensitivity of the algorithm to the different parameters used in the method is explored. The effect of noise on the proposed method is also tested, and finally we compare our results to other well-established methods from the research literature. The two main contributions of this work are: first, demonstrat- ing that a stochastic, distributed, agent-based, and swarm-based method can perform an image processing task such as feature extraction with high accuracy, outperforming traditional methods, even in the presence of noise, and second, developing a new algo- rithm for image edge extraction which can be tuned using different parameters for satisfying performance in the presence of noise. In Section 2 of this paper we review some of the related work in this area. In Section 3 an overview of the problem central to this paper is presented. A brief definition of the problem is first described in Section 3.1, and an overview of ant-based swarm intelligence is reviewed in Section 3.2. Subsequently, the proposed model is described in Section 4 and details of the methodology are described throughout this section. Section 5 explains the post- processing procedures required subsequent to execution of the main algorithm, as well as the methods employed for evaluation of the proposed feature extraction system. In Section 6 the experi- mental setup and respective results for validation of the proposed method are described and presented, and accordingly, respective discussions are provided. Finally, in Section 7 conclusions and final remarks are presented. 2. Related work In the field of image processing, the use of soft computing, and ant-based swarm intelligence in particular, do not have a long 1568-4946/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.asoc.2011.06.011