A Deep Learning Approach For Object
Detection
Problem Background
Object detection is a subfield of computer vision that is currently heavily based on
machine learning. For the past decade, the field of machine learning has been
dominated by so called deep neural networks, which take advantage of
improvements in computing power and data availability. A subtype of a neural
network called a convolutional neural network (CNN) is well-suited for image-
related tasks. The network is trained to look for different features, such as edges,
corners and colour differences, across the image and to combine these into more
complex shapes. For object detection, the system has to both estimate the
locations of probable objects and to classify these.
Problem & Objectives
Problem Statement:
• To study and evaluate Faster R-CNN approach for object detection.
• To build a deep neural network that will detect 12 object classes from an image.
Objectives:
• To develop an object detection and classification method.
• To detect and classify multiple objects using a single image.
Methodology
Data Data Define
Preparation Augmentation Parameters
Train the
Optimization Define Model
model
Testing and
Calculate loss
Evaluation
Figure 1
Architecture of Modified Faster R-CNN
Sample Results
Parameters Faster R-CNN Modified Faster R-CNN
Mean Average
Techniques Precision
Class Accuracy 82.6 85
RPN class loss 0.1 0.08 Faster R-CNN using VGG-16 37.2
RPN regression loss 0.16 0.15
Faster R-CNN using ResNet-50 34.1
Classification class loss 0.42 0.38
Faster R-CNN using Xception 33.8
Classification regression loss 0.21 0.18
Faster R-CNN using Inception ResNet v2. 34.3
Total loss 0.87 0.75
Modified Faster R-CNN using VGG-16 44.3
Mean average precision 37.2 44.3
Table 1 Table 2
Sample Results
Figure 3
Sample Results
Figure 4
Sample Results
Figure 5
Research Interest
• Deep Learning
• Cloud computing
• Internet of Things
THANK YOU
