Edge detection of infrared image with CNN_DGA algorithm

doi:10.1016/j.ijleo.2013.07.049

Optik

Volume 125, Issue 1, January 2014, Pages 464-467

https://doi.org/10.1016/j.ijleo.2013.07.049 Get rights and content

Abstract

In this paper, infrared image edge detection algorithm based on the combination of the cellular neural networks (CNN) and distributed genetic algorithm (DGA) is proposed. The CNN template is used to train the network with distributed genetic algorithm (CNN_DGA). The experimental results show: the edge of the infrared image was extracted accurately with CNN_DGA edge detection algorithm; furthermore, the noise of the infrared image was reduced greatly. Compared with the edge detection algorithms based on cellular neural networks with template trained by particle swarm optimization, parameters’ search range and convergence speed are greatly improved.

Introduction

The edge detection technology is usually used to extract of edge feature of the image. Edge feature is one of the most fundamental and important feature of image and it can be used to represent the image. In recent years, many new theories are proposed in image processing area, such as cellular neural network, genetic algorithm, wavelet transform, particle swarm algorithm [1] and clone selection algorithm [2], more and more attention were focused on the research of the algorithms for processing of image.

The brightness intensities of infrared images are representative of the temperature of object surface. Image intensities are clustered in small regions in image intensity histogram, while intensity histogram for typical visible images is close to constant for better utilizing the dynamic range of intensity value. Compared with rich and colorful visible images, infrared images are blurrier, have poorer resolution and clarity, and foreground/background contrast is less clear [3]. So, there are some difficulties to deal with infrared image with traditional image processing algorithms.

The main edge detection algorithms are based on operators, such as Sobel operator, Roberts operator, Prewitt operator and Canny operator [4]. But unfortunately, they are not so suitable for the edge detection of infrared images. The application of the new theories gives an opportunity for finding new method of edge detection. In cellular neural networks (CNN) algorithm [5], the linear and nonlinear circuit elements which processes signals in real time are included. The CNN have some circuit clones, called cells, which communicate each of its nearest neighbors [6]. The cells represent the gray value of pixel in infrared image.

Although S. Schwarz et al. proposed a learning algorithm for CNN in [7] and [8], the CNN has the restriction of the smallest neighborhood. The distributed genetic algorithm (DGA) is proposed to overcome the problems of stability and adaptation in the CNN. CNN template is selected by DGA, and it is known to be powerful and robust means [9], [10]. Due to the good searching capacity of CNN, the CNN-DGA combined edge detection algorithm for infrared image is proposed in this paper.

This paper is organized as follows. Section II describes the principle of cellular neural network. In section III, we introduce the design of edge detection template based on DGA. And Section IV, some illustrative examples is analyzed with proposed CNN-DGA edge detection algorithm. The conclusions are drawn in the last section.

Section snippets

Cellular neural network

The cellular neural network (CNN) is a locally connected parallel information processing system [11]. It contains linear and nonlinear circuit elements, which typically are linear capacitors, linear resistors, linear and nonlinear controlled sources, and independent sources [12]. For the infrared image processed by CNN, the cell dynamical equation as follow: $C \frac{d x_{i j} (t)}{d t} = - \frac{x_{i j} (t)}{R_{x}} + \sum_{k, l \in N_{i j} (r)} A_{k l} y_{k l} (t) + \sum_{k, l \in N_{i j} (r)} B_{k l} u_{k l} + I_{i j}$ where A_kl and B_kl are element of feedback template and feed-forward

Experimental results and discussion

In this section, the edge detecting of the infrared image with the proposed algorithm is illustrated. The original infrared image used for simulation is shown in Fig. 4. There are a number of parameters in a genetic algorithm which have to be specified. The parameters used in the simulation are as follow: population size is 30, initial mutation rate is 0.2, iteration cycle is 50, initial crossover rate is 0.8, sub-populations number is 3, and the population of the initial individual is ranged

Conclusion

The edge detection algorithm is useful to detect the precise edge for infrared images. The proposed CNN_DGA edge detection algorithm can get clear edge information for infrared image even if the signal to noise ratio is low and the original infrared images without any preprocessing. Both the convergence time and the search range of the parameters are greatly improved.

References (15)

H.A. Firpi et al.
Designing templates for cellular neural networks using particle swarm optimization
ISIT
(2004)
A. Basturk et al.
Clonal selection algorithm based cloning template learning for edge detection in digital images with CNN
IEEE Conf. Sig. Process. Commun. Appl.
(2008)
F. Dongzhu et al.
An algorithm of infrared imaging edge detection based on grey relational analysis
Infrared Technol.
(2006)
C.G. Rafael et al.
Digital Image Processing Using MATLAB
(2005)
J.J. Hopfield
Neural networks and physical systems with emergent computational abilities
Proc. Natl. Acad. Sci. U.S.A.
(1982)
L.O. Chua et al.
Cellular neural networks: applications
IEEE Trans. Circ. Syst.
(1988)
F. Zou et al.
Cellular neural network design using a learning algorithm

There are more references available in the full text version of this article.

Cited by (19)

A digitalization-based image edge detection algorithm in intelligent recognition of 5G smart grid
2023, Expert Systems with Applications
In the 5G smart grid, battery liquid leakage negatively affects the safe storage and transportation of electricity. Image edge detection help locate the liquid leakage phenomenon. This paper proposes a digitalization-based algorithm that converts the image to the codeword and judges the edge by the code weight. Its main steps are image sampling, quantization, coding and judging. Compared with traditional Sobel and Canny operators, the algorithm dramatically enhances detection ability with machine learning. Moreover, this paper applies the algorithm to detect the leakage of infrared battery images in the 5G smart grid; The results show that it can accurately extract the liquid leakage location, improving fault diagnosis and early risk warning.
Benign and malignant breast tumors classification based on region growing and CNN segmentation
2015, Expert Systems with Applications
Citation Excerpt :
Wang et al. proposed CNN learning algorithms based on genetic algorithm and particle swarm optimization (PSO) for infrared images. Based on their results, genetic algorithm showed fast convergence compared to PSO, especially for large images (Wang, Yang, Xie, & An, 2014). In fact, in the second proposed technique, the CNN parameters are set using a genetic algorithm with an attempt to preserve tumor boundary in mammograms segmentation, specially for the images in DDSM and MIAS databases.
Breast cancer is regarded as one of the most frequent mortality causes among women. As early detection of breast cancer increases the survival chance, creation of a system to diagnose suspicious masses in mammograms is important. In this paper, two automated methods are presented to diagnose mass types of benign and malignant in mammograms. In the first proposed method, segmentation is done using an automated region growing whose threshold is obtained by a trained artificial neural network (ANN). In the second proposed method, segmentation is performed by a cellular neural network (CNN) whose parameters are determined by a genetic algorithm (GA). Intensity, textural, and shape features are extracted from segmented tumors. GA is used to select appropriate features from the set of extracted features. In the next stage, ANNs are used to classify the mammograms as benign or malignant. To evaluate the performance of the proposed methods different classifiers (such as random forest, naïve Bayes, SVM, and KNN) are used. Results of the proposed techniques performed on MIAS and DDSM databases are promising. The obtained sensitivity, specificity, and accuracy rates are 96.87%, 95.94%, and 96.47%, respectively.
Video detection algorithm based on semantic segmentation for conveyor belt deviation
2023, China Safety Science Journal
Accurate segmentation of infrared images for circuit board diagnosis using an improved Deeplabv3+ network
2023, Frontiers in Physics
GPNRBNN: A Robot Image Edge Detection Method Based on Gaussian Positive-Negative Radial Basis Neural Network
2021, Sensing and Imaging
A hybrid artificial bee colony with whale optimization algorithm for improved breast cancer diagnosis
2021, Neural Computing and Applications

View all citing articles on Scopus

View full text

Edge detection of infrared image with CNN_DGA algorithm

Abstract

Introduction

Section snippets

Cellular neural network

Experimental results and discussion

Conclusion

Designing templates for cellular neural networks using particle swarm optimization

ISIT

Clonal selection algorithm based cloning template learning for edge detection in digital images with CNN

IEEE Conf. Sig. Process. Commun. Appl.

An algorithm of infrared imaging edge detection based on grey relational analysis

Infrared Technol.

Digital Image Processing Using MATLAB

Neural networks and physical systems with emergent computational abilities

Proc. Natl. Acad. Sci. U.S.A.

Cellular neural networks: applications

IEEE Trans. Circ. Syst.

Cellular neural network design using a learning algorithm