IJSRSET1848148 | Received : 15 June 2018 | Accepted : 26 June 2018 | May-June-2018 [(4) 8 : 640-645]
© 2018 IJSRSET | Volume 4 | Issue 8 | Print ISSN: 2395-1990 | Online ISSN : 2394-4099
Themed Section : Engineering and Technology
640
An Efficient Algorithm for Real-Time Object Detection in Images
Md. Samiul Islam, Samia Sultana
Stamford University Bangladesh, Dhaka, Bangladesh
ABSTRACT
In this paper we have proposed an algorithm for object detection in various situations. Nowadays object
detection and recognition has entered in every sphere of life in one or the other form. Applications of object
detection are video surveillance, anti-theft system using cameras, face-recognition, biometric verification etc.
Research are going on how to improve the performance in term of space and time complexity, how to deal with
adverse conditions like improper lightning conditions, scene clutter, occlusion etc. and to reduce false positive
rate etc. In this paper we have explained how to deal with the any situation while acquiring the images so that
it can be used for better scene interpretation. Results have been generated using flash of light and dark region
present in the image as some of the adverse situations. Here we have trained the system to detect the object
using our algorithm. The algorithm is simple and very useful as it reduces the false positive rate as compared to
contemporary algorithms and increases the efficiency of applications like video surveillance and scene
interpretation etc.
Keywords: Object Detection, Image Classification, Image Recognition, Histogram of Oriented Gradients,
Support Vector Machine.
I. INTRODUCTION
An image recognition algorithm takes an image (or a
patch of an image) as input and outputs what the
image contains. In other words, the output is a class
label (e.g. “cat”, “dog”, “table” etc.). How does an
image recognition algorithm know the contents of an
image? Well, we have to train the algorithm to learn
the differences between different classes. If we want
to find cats in images, we need to train an image
recognition algorithm with thousands of images of
cats and thousands of images of backgrounds that do
not contain cats. Needless to say, this algorithm can
only understand objects / classes it has learned.
To simplify things, we will focus only on two-class
(binary) classifiers. One may think that this is a very
limiting assumption, but many popular object
detectors (e.g. face detector and pedestrian detector)
have a binary classifier under the hood. E.g. inside a
face detector is an image classifier that says whether a
patch of an image is a face or background.
The following diagram illustrates the steps involved
in a traditional image classifier.
Figure 1: Object (cat) detection using traditional
image processing
Interestingly, many traditional computer vision image
classification algorithms follow this pipeline, while