IMAGE RECOGNITION USING MATLAB SIMULINK BLOCKSET

International Journal of Computer Science, Engineering and Applications (IJCSEA) Vol. 7, No. 2, April 2017
1
IMAGE RECOGNITION USING MATLAB SIMULINK
BLOCKSET
Sagir Lawan and CL Wamdeo
Department of Electrical/Electronic Engineering, Nigerian Defence Academy, Kaduna
ABSTRACT
The world over, image recognition are essential players in promoting quality object recognition especially
in emergency and search-rescue operation. In this paper precise image recognition system using Matlab
Simulink Blockset to detect selected object from crowd is presented. The process involves extracting object
features and then recognizes it considering illumination, direction and pose. A Simulink model has been
developed to eliminate the tiny elements from the image, then creating segments for precise object
recognition. Furthermore, the simulation explores image recognition from the coloured and gray-scale
images through image processing techniques in Simulink environment. The tool employed for computation
and simulation is the Matlab image processing blockset. The process comprises morphological operation
method which is effective for captured images and video. The results of extensive simulations indicate that
this method is suitable for application identifying a person from a crow. The model can be used in
emergency and search-rescue operation as well as in medicine, information security, access control, law
enforcement, surveillance system, microscopy etc.
KEYWORDS
Matlab, Morphological Operations, identification, Target Recognition, Simulink, Video and Image
Processing
1. INTRODUCTION
Presently, there exists a good number of software for digital image processing in object
recognition. Image recognition has varied applications in medicine, information security, access
control, law enforcement, surveillance system, microscopy etc. Image processing techniques is
used to modify pictures to improve them for recognition. In some cases image processing is used
to change images structure by optical, photographic and electronic means. Image processing
using digital computers is the most common because digital methods are fast, flexible and
precise. Hence, computer algorithms are continuously being developed to achieve better results.
The search for more reliable, cheaper and less complex computation led to use of Matlab
Simulink Model Blockset for image recognition.
In making operational decisions especially in the military operations in urban terrain, it may be
possible to take into consideration image processing to aid in object recognition. This paper aims
at identifying important object in images captured from scene using Blockset in Simulink
environment of MATLAB. The Video and Image Processing Blockset contains blocks that
perform morphological operations such as erosion, dilation, opening and closing. A Simulink
model has been developed to eliminate the tiny elements from the image, then creating segments
for precise object recognition [1]. The remainder of the paper is organized as follows. Section 2
briefly presents the related research. Section 3 briefly describes on various types of object
recognition techniques. Section 4 presents experimental environment. Section 5 covers the
results and analysis. Finally, Section 5 draws the conclusion.

2
2. RELATED RESEARCH
The idea of target identification has taken various forms prior to image processing. A
considerable amount of literature has also been published on digital image processing and
intelligent target recognition. In [2], an approach on image/video segmentation was taken in
which existing video object segmentation algorithm and morphological operation was used to
analyse core operations. Hence, morphology image processing element array and stream
processing was used. The author, further stated that control and interconnection between
processing element can also be configured. This view is supported by [3], the authors experiment
with FPGA using Xlinx system generator to implement MRI image filtering and tour
characterization. They focused on processing an image pixel by pixel and in modification of pixel
neighbourhoods and the transformation that can be applied to the whole image or partial region.
This work was complemented by [4] study of pixel to pixel processing of an image and the
modification of pixel transformation to the whole image or only a partial region. However, a big
challenge to this approach is feature extraction.
More recent attention has focused on the problem of recognizing of vehicles on a crowed road as
contained in [1]. The author found that vehicle counting on roads can reliably be performed using
traffic load computation with Matlab Simulink Model Blockset. They stated that Simulink model
can be used to detect a particular vehicle in traffic on the road. Their algorithm proves to be an
efficient solution for computing traffic load. A broader perspective has been adopted by [5] who
argues that segmentation algorithms based on one of two basic properties partition an image
based on abrupt change in intensity and regions that are similar according to predefined criteria.
Although this is the most comprehensive Simulink image processing, the methods lacks clear
details for identification of target. In the same vein, [6] noted the detection and tracking of human
as well as vehicles. The paper hitherto have attempted to draw fine distinctions between detection
and tracking in three ways (i) automatic target detection and recognition tracking technique, (ii)
tracking object across a network and (iii) development of real time communication and message
networks.
[7 ] introduced a novel optical encryption and decryption method of gray image based on the
computer-generated hologram (CGH). However, in the encryption process, the hologram, which
is gotten by Fresnel diffraction from the gray image, is modulated with the chaotic sequence, and
the output is encryption image. According to the author in [8] image recognition was achieved
through noise reconstruction of holographic images in the form of a system of fringes parallel to
the hologram frame boundaries. The author assert that mathematical and physical interpretation is
proposed together with an algorithm for reduction of image effect using bicubic splines. Trofimov
el at, [9] proposed an improved and a real-time algorithm of computer processing for
image recognition using passive THz image with the aim of concealed object detection.
They stated that the algorithm used correlation function between the THz image and a
standard image. The result detect the presence of forbidden objects on the human body..
3. OBJECT RECOGNITION METHODS
In the 1970's, object recognition was treated as a 2D pattern recognition problem [2]. The
distances between important points where used to recognize known object, however, with
technological drive object 2D and 3D objects recognition methods are fully automated. Object
recognition methods include holistic matching methods, feature-based (structural) methods and
hybrid methods are widely used.

3
In both 2D and 3D holistic approach, the complete object region is taken into account as input
data into catching system. One of the best examples of holistic methods is Eigenfaces [8] (most
widely used method for face recognition). This approach covers face recognition as a two-
dimensional recognition problem. The process involves insertion of a set of images into a
database to create the eigenfaces. Eigenfaces of humans are made by extracting characteristic
features from the faces. The input images are normalized to line up the eyes and mouths. When
the eigenfaces have been created, each image will be represented as a vector of weights. The
weight of the incoming unknown image is found and then compared the weights of those already
in the system. If the input image's weight is over a given threshold it is considered to be
unidentified. [2].
Another method is described as feature-based (structural) methods: In this methods local features
such as eyes, nose and mouth are first of all extracted and their locations and local statistics
(geometric and/or appearance) are fed into a structural classifier. A draw back to this method is
feature extraction and image restoration [5]. In compliance with need for accurate image
recognition, a hybrid method was developed. This method recognized face in both 2D and 3D
pattern using a combination of both holistic and feature extraction methods.
4. EXPERIMENTAL ENVIRONMENT
To date various models for image processing have been developed and introduced. In this paper
hybrid method model based design is configured to identify and recognize important object from
a crowd using Matlab Simulink Model. The process involves use of Matlab numerical computing
environment and matrix manipulations. Simulink model is used to add graphical multi domain
simulation and embedded systems. Signal processing is achieved by adapting to the existing
algorithms contained in MATLAB programs. Figure 1 shows block diagram of simulation
process.
Figure 1: Simulink Model
The basic model design and the fundamental work to read and write images through Simulink is
shown in Figures 2 – 9. The intensity of the image under experiment is imported From Simulink
library browser to create a new Simulink model as indicated in Figure 2.

4
Figure 2: Simulink Library
Figure 3: New Model
The values of the block parameters are modified to complain with the specification of the targeted
object as demonstrated in Figures 4, 5 and 6. The constant value parameter is set to 200, this
defined the threshold. The Rational operator is set to” < ” and the configuration parameters is set
to solver pane stop time to zero, solver tyre to fixed-step and solver to discrete (no continuous
state). The simulated image workspace is illustrated in Figures 7, 10, 13, 16 and19.

5
Fig 4: Threshold
Figure 5: Opening Block

6
Figure 6: Label Block
5. RESULTS
This section describes the results of the simulations conducted; the overall result is presented in
terms of: A. coloured and gray-scale image, B. identification and C. boundary performance.
A. Coloured and Gray-Scale Images. The conversion of colour to gray process is done by
making all pixels above a certain threshold level white while, others black. The suitable threshold
constant taken is 200. Figures 8, 11, 14, and 17 shows results of the image inputs been converted
to gray.
Figure 7: Input Image “Classroom”

7
Figure 8: Classroom “Gray”
Figure 9: Classroom Object Identified
B. Identification. Identification here is simply masking operation and the main function in the
opening block create a circular STREL object with radius of 5 pixels that separate the important
objects from the image and from each other at the image. The results of the objects identified in
gray are depicted in Figures 9, 12, 15 and 18.

8
Figure 10: Input image “River side”
Figure 11: River side Gray
Figure 12: River side Object Identified
C. Boundary. For boundary identification, the image is filtered and masked to ensure better noise
suppression and smoothing. The STREL function provided this as illustrated in Figures 9, 12, 15
and 18.

9
Figure 13: Input image “Youth”
Figure 14: Youth Gray

10
Figure 15: Youth Object Identified
Figure 16: Input image “Lab”
Figure 17: Lab Gray
Figure 18: Lab Object Identified

11
6. CONCLUSION
In this paper, we presented an image recognition using Matlab Smulink blockset that facilitates
flexible identification and recognition of objet in a crowd. Our goal is to separate identified object
from many based on configuration and content management of Matlab Smulink blockset. When
employing hybrid object identification method for 2D and 3D patterns, the algorithm is easily
updated at any time. The proposed method allows reliable and dynamic object identification that
will lead to making critical decisions during emergency and search-rescue operations. Feature
steps include research on facial feature extraction of human face. Another aspect is integration of
the proposed model towards object projection of face images into a linear subspace by reducing
dimension.
REFERENCES
[1] Pratishtha G, Purohit G.N, Adhyana Gupta, “Traffic Load Computer using Matlab Simulink Model
Blockset” international Journal of Advanced Research in Computer and Communication Engineering
Vol. 2, Issue6, June 2013.
[2] Chien S.Y. ,Chen L.G., “Reconfigurable Morphological Image Processing Accelerator for Video
Object Segmentation”, springer, J Sign Process Syst (2011) 62:77–96, DOI 10.1007/s11265-008-
0311-6
[3] Christe S.A., Vignesh M., Kandaswamy A., “An efficient FPGA implementation of MRI image
filtering and tumour characterization using Xilinx system generator”, International Journal of VLSI
design & Communication Systems (VLSICS), Vol.2 ,No.4, December 2011
[4] Suthar A. C., Vayada M., Patel C. B., Kulkarni G.R., Hardware Software co-simulation for Image
Processing Applications”, IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 2, No
2, March 2012
[5] Al-amri S.S., Kalyankar N.V., Khamitkar S.D, “Image Segmentation by Using Threshold
Techniques”, journal of computing, volume 2, issue 5, may 2010, ISSN 2151-9617 [6]Shinde B.,
Mhaske D., Dani A.R., “Study of Image Processing, Enhancement and Restoration”, IJCSI
International Journal of Computer Science Issues, Vol. 8, Issue 6, No 3, November 2011
[7] Mahalanobis A., Cannon J., Stanfill S.R., Muise R., Martin L., “Network Video Image Processing for
Security, Surveillance, and Situational Awareness”, Proc. SPIE 5440, Digital Wireless
Communications VI,1 (August 10, 2004); doi:10.1117/12.548981
[8] Ren, Hui, Jun Wang, and Qiong-hua Wang. "Optical Encryption of Gray Image Based on the
Computer Generated Hologram and Logical Modulation." Proc. of SPIE Vol. Vol. 10250. 2017.
[9] Dyomin, V., and A. Olshukov. "Improvement of the Quality of Reconstructed Holographic Images by
Extrapolation of Digital Holograms." Russian Physics Journal 58.10 (2016).
[10] Trofimov, Vyacheslav A., et al. "Concealed object detection using the passive THz image without its
viewing." SPIE Defense+ Security. International Society for Optics and Photonics, 2016.

IMAGE RECOGNITION USING MATLAB SIMULINK BLOCKSET

More Related Content

What's hot

Similar to IMAGE RECOGNITION USING MATLAB SIMULINK BLOCKSET

Recently uploaded

IMAGE RECOGNITION USING MATLAB SIMULINK BLOCKSET