Abstract:
The main objective ofthis project is to localize and detect the any moving object in a
real time video and obviously that can be one ofthe challenging task in the fields like
the computer vision due to several factors such as the centroid or the origin of the
object, the location ofthe object different types ofshadings in a video to detect etc.
By studying the Many ofthe Artificial Intelligence algorithm one ofthe most
efficient that we have found is the Convolutional Neural Network or the CNN that
has the good accuracy rate and yet the most perfect one to fit for recognizing any
custom trained object in a video, A Module will be developed which will further
train and detect the custom objects and all the final implementation will be done
using the Programming language called ‘Python5 IDE (PYcharm).
Just to train a single object there the minimum requirement for a single object
than the 900 images and if increase the system will get more efficient in
detecting that image and by considering that I have noticed that Convolutional
Neural Network is the better choice than all other deep learning algorithms there
would be one of the training phase and the testing phase as well. Further
is more
development is also included in the report.
Object Detection is the process of finding real-world object instances like
y real life time objects still images or Videos. It allows for the recongnition ,
localization, and detection ofmultiple objects within an image which provide us with
h understanding of an image as a whole. It is commonly used in applications
retrieval, security, surveillance, and advanced driver assistance system
man
a muc
such as image
(ADAS).
We work on different data set and train them to detect an objects , it give us
100% accuracy result through CNN algorithm.
