Authors :
Manish Yewange; Kanishk Gaikwad; Ramkishan Kamble; Sagar Maske; D.T.Salunke
Volume/Issue :
Volume 7 - 2022, Issue 4 - April
Google Scholar :
https://bit.ly/3IIfn9N
Scribd :
https://bit.ly/3LzfZzO
DOI :
https://doi.org/10.5281/zenodo.6568421
Abstract :
Object detection has received a significant
attention from researchers in recent years because of its
close relationship with video analysis and image
understanding. Traditional object detection approaches
are built on the basis of classifiers and shallow trainable
architectures. Their performance can easily plateau by
developing sophisticated ensembles that incorporate
various low-level visual features with high-level context
from object detectors and scene classifiers. More
powerful tools that can learn semantic, high-level, and
deeper features are being offered with the rapid growth
of deep learning to address the difficulties that afflict
traditional systems. These models react differently in
terms of network architecture, training approach, and
optimization function. We present a review of deep
learning-based object detection frameworks in this
paper. Object recognition tasks, on the other hand, are
more difficult to assess, consume more energy, and
necessitate more computing power than image
categorization. To address these issues, a novel technique
for real-time object detection application is being
developed in order to improve the detection process'
accuracy and energy efficiency.
Keywords :
CNN, F-RCNN, Dataset, Images, Videos.
Object detection has received a significant
attention from researchers in recent years because of its
close relationship with video analysis and image
understanding. Traditional object detection approaches
are built on the basis of classifiers and shallow trainable
architectures. Their performance can easily plateau by
developing sophisticated ensembles that incorporate
various low-level visual features with high-level context
from object detectors and scene classifiers. More
powerful tools that can learn semantic, high-level, and
deeper features are being offered with the rapid growth
of deep learning to address the difficulties that afflict
traditional systems. These models react differently in
terms of network architecture, training approach, and
optimization function. We present a review of deep
learning-based object detection frameworks in this
paper. Object recognition tasks, on the other hand, are
more difficult to assess, consume more energy, and
necessitate more computing power than image
categorization. To address these issues, a novel technique
for real-time object detection application is being
developed in order to improve the detection process'
accuracy and energy efficiency.
Keywords :
CNN, F-RCNN, Dataset, Images, Videos.
