Chapter 5: Construction

An object detection technique lets you understand the details of an image or a video as it allows for
the recognition, localization, and detection of multiple objects within an image.
It is usually utilized in applications like image retrieval, security, surveillance, and advanced driver
assistance systems (ADAS). Object Detection is done through many ways:

 Feature Based Object Detection

 Viola Jones Object Detection
 SVM Classifications with HOG Features
 Deep Learning Object Detection

DIGITAL IMAGE PROCESSING

Computerized picture preparing is a range portrayed by the requirement for broad test work to build
up the practicality of proposed answers for a given issue. A critical trademark hidden the plan of
picture preparing frameworks is the huge level of testing and experimentation that

Typically, is required before touching base at a satisfactory arrangement. This trademark informs
that the capacity to plan approaches and rapidly model hopeful arrangements by and large assumes
a noteworthy part in diminishing the cost and time required to land at a suitable framework
execution.

Processing on image:

Processing on image can be of three types They are low-level, mid-level, high level.

Low-level Processing:

 Preprocessing to remove noise.

 Contrast enhancement.

 Image sharpening.

Medium Level Processing:

 Segmentation.

 Edge detection

 Object extraction.

High Level Processing:

 Image analysis
 Scene interpretation

YOLO: Real-Time Object Detection

You only look once (YOLO) is a state-of-the-art, real-time object detection system. On a Pascal Titan
X it processes images at 30 FPS and has a mAP of 57.9% on COCO test-dev.

Comparison to Other Detectors

YOLOv3 is extremely fast and accurate. In mAP measured at .5 IOU YOLOv3 is on par with Focal Loss
but about 4x faster. Moreover, you can easily tradeoff between speed and accuracy simply by
changing the size of the model, no retraining required!
Performance on the COCO Dataset

Model Train Test mAP FLOPS FPS

SSD300 COCO trainval test-dev 41.2 - 46

SSD500 COCO trainval test-dev 46.5 - 19

YOLOv2 608x608 COCO trainval test-dev 48.1 62.94 Bn 40

Tiny YOLO COCO trainval test-dev 23.7 5.41 Bn 244

SSD321 COCO trainval test-dev 45.4 - 16

DSSD321 COCO trainval test-dev 46.1 - 12

R-FCN COCO trainval test-dev 51.9 - 12

SSD513 COCO trainval test-dev 50.4 - 8

DSSD513 COCO trainval test-dev 53.3 - 6

FPN FRCN COCO trainval test-dev 59.1 - 6

Retinanet-50-500 COCO trainval test-dev 50.9 - 14

Retinanet-101-500 COCO trainval test-dev 53.1 - 11

Retinanet-101-800 COCO trainval test-dev 57.5 - 5

YOLOv3-320 COCO trainval test-dev 51.5 38.97 Bn 45

YOLOv3-416 COCO trainval test-dev 55.3 65.86 Bn 35

YOLOv3-608 COCO trainval test-dev 57.9 140.69 Bn 20

YOLOv3-tiny COCO trainval test-dev 33.1 5.56 Bn 220

YOLOv3-spp COCO trainval test-dev 60.6 141.45 Bn 20

How It Works
Prior detection systems repurpose classifiers or localizers to perform detection. They apply the
model to an image at multiple locations and scales. High scoring regions of the image are considered
detections.

We use a totally different approach. We apply a single neural network to the full image. This
network divides the image into regions and predicts bounding boxes and probabilities for each
region. These bounding boxes are weighted by the predicted probabilities. Our model has several
advantages over classifier-based systems. It looks at the whole image at test time so its predictions
are informed by global context in the image. It also makes predictions with a single network
evaluation unlike systems like R-CNN which require thousands for a single image. This makes it
extremely fast, more than 1000x faster than R-CNN and 100x faster than Fast R-CNN. See our paper
for more details on the full system.

What's New in Version 3?

YOLOv3 uses a few tricks to improve training and increase performance, including: multi-scale
predictions, a better backbone classifier, and more. The full details are in our paper!

Detection Using a Pre-Trained Model

This post will guide you through detecting objects with the YOLO system using a pre-trained model.
If you don't already have Darknet installed, you should do that first.

Algorithm:

Yolo

Step 1: if (setModel=YoloV3( )|| TinyYoloV3( ))

Step 2: Set the execution path

Step 3: Set the model path to load the model

Step 4: Set timer as default timer to get the elapsed time

Step 5: Detect the objects from input to output image

Step 6: join the detector on inputImage and outputImage

Step 7: obj:=0, accur:=0

Step 8: for each object in detection do

Step 9: {

Step 10 :obj:=obj+1; accr:=accr+percentprob;

Step 11: }

Step 12: eTime: = defaultTimer – startTime;

5.1 Implementation and testing:

Real-Time Detection on a Webcam

Running YOLO on test data isn't very interesting if you can't see the result. Instead of running it on a
bunch of images let's run it on the input from a webcam!

To run this demo, you will need to compile Darknet with CUDA and OpenCV. Then run the
command:

./darknet detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights

YOLO will display the current FPS and predicted classes as well as the image with bounding boxes
drawn on top of it.

You will need a webcam connected to the computer that OpenCV can connect to or it won't work. If
you have multiple webcams connected and want to select which one to use you can pass the flag -c
<num> to pick (OpenCV uses webcam 0 by default).

You can also run it on a video file if OpenCV can read the video:

./darknet detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights <video file>

Implemented Classes:

Some of the main classes that have been implemented in the project are as follows
there are many kind of classes that are present in the weights section of the YOLOv3
but some of the main classes are as follows.

ClassIndex: Used for defining the index of the particular class and the array that we have
created using the coco dataset in our project.

Confidence: Used for creating the boxes outside of the object that has been detected using
the YOLO.It can be of green or any color we want.
Bbox: Used for creating the boxes outside of the object that has been detected using the
YOLO. It can be of green or any color we want.

Implemented Functions

Weighted Sum

Inputs to a neuron can either be features from a training set or outputs from the neurons of
a previous layer. Each connection between two neurons has a unique synapse with a unique
weight attached. If you want to get from one neuron to the next, you have to travel along
the synapse and pay the “toll” (weight). The neuron then applies an activation function to
the sum of the weighted inputs from each incoming synapse. It passes the result on to all the
neurons in the next layer. When we talk about updating weights in a network, we’re talking
about adjusting the weights on these synapses.

A neuron’s input is the sum of weighted outputs from all the neurons in the previous layer.
Each input is multiplied by the weight associated with the synapse connecting the input to
the current neuron. If there are 3 inputs or neurons in the previous layer, each neuron in the
current layer will have 3 distinct weights: one for each synapse.

In a nutshell, the activation function of a node defines the output of that node.

The activation function (or transfer function) translates the input signals to output signals. It
maps the output values on a range like 0 to 1 or -1 to 1. It’s an abstraction that represents
the rate of action potential firing in the cell. It’s a number that represents the likelihood that
the cell will fire. At it’s simplest, the function is binary: yes (the neuron fires) or no (the
neuron doesn’t fire). The output can be either 0 or 1 (on/off or yes/no), or it can be
anywhere in a range. If you were using a function that maps a range between 0 and 1 to
determine the likelihood that an image is a cat, for example, an output of 0.9 would show a
90% probability that your image is, in fact, a cat.

Activation function

In a nutshell, the activation function of a node defines the output of that node.

The activation function (or transfer function) translates the input signals to output signals. It
maps the output values on a range like 0 to 1 or -1 to 1. It’s an abstraction that represents
the rate of action potential firing in the cell. It’s a number that represents the likelihood that
the cell will fire. At it’s simplest, the function is binary: yes (the neuron fires) or no (the
neuron doesn’t fire). The output can be either 0 or 1 (on/off or yes/no), or it can be
anywhere in a range.
 Threshold function

This is a step function. If the summed value of the input reaches a certain threshold the
function passes on 0. If it’s equal to or more than zero, then it would pass on 1. It’s a very
rigid, straightforward, yes or no function.

Sigmoid function

This function is used in logistic regression. Unlike the threshold function, it’s a smooth,
gradual progression from 0 to 1. It’s useful in the output layer and is used heavily for linear
regression.

Hyperbolic Tangent Function

This function is very similar to the sigmoid function. But unlike the sigmoid function which
goes from 0 to 1, the value goes below zero, from -1 to 1. Even though this isn’t a lot like
what happens in a brain, this function gives better results when it comes to training neural
networks. Neural networks sometimes get “stuck” during training with the sigmoid function.
This happens when there’s a lot of strongly negative input that keeps the output near zero,
which messes with the learning process.

Rectifier function

This might be the most popular activation function in the universe of neural networks. It’s
the most efficient and biologically plausible. Even though it has a kink, it’s smooth and
gradual after the kink at. This means, for example, that your output would be either “no” or
a percentage of “yes.” This function doesn’t require normalization or other complicated
calculations.

It comes under the layer of machine learning, where machines can acquire skills and learn
from past experience without any involvement of human. Deep learning comes under
machine learning where artificial neural networks, algorithms inspired by the human brain,
learn from large amounts of data.

4.2 Validation and Verification

Although Objection detection is an esteemed task yet, it is an innovative errand. It plays an essential
role in numerous implementations like identifying an image, auto-annotation of image, and
apprehension of the ideology. Eliminating the problem of vision in visually impaired persons, the
proposed work can be used effectively in detecting the objects along with their design patterns in an
exact manner and to identify them among multiple different objects in a captured input image
individually with high accuracy and with expert navigation, by implementing the Specific model X-Y
plane by calculating their percentages accurately of the detection and also supporting the
transformation input images to speech. The object detection also furnishes its results on multiple
objects and various methodologies in discovering artefacts, identifying and collating each step for its
productiveness.

4.3 Testing Approach:

The image or video can be loaded into the object detection model, this interface contains loading an
image, running a module to execute the program, the number of detected images detected in the
module and play audio for better understanding for visually impaired persons.

Figure 7: Interface to run the detection modules.

Figure 8: Object detection in an outdoor environment with multi labelling.

Figure 8 shows the loaded image in the outdoor environment at one part, and the other hands, the
model is marked with all the objects available in the picture with blue-coloured frames.

Figure 9: Accuracy values of available objects in the image. Figure 9 shows all the available objects in
the images with accuracies. The detection module observed that there are five bottles, one chair,
and eight persons are detected in the loaded image. By playing the audio, the module says, “Hey!
There are five bottles, one chair’s eight people’s before you”.
Figure 10: Object detection in a traffic environment.

Figure 10 shows all the available variables with labels at the traffic signal environment.

In this case, the model detected seven cars, two trucks, one person’s, and one bicycle’s before the
person, along with accuracy.

Figure 11: Playing audio output. Figure 11 shows the accuracy of the objects available in the loaded
image. By playing the “play audio” module, the visually impaired people can listen to the type of
objects in the surrounding environment and the count.

Figure 12: Object detection at a traffic signal.

Figure 12 shows all the detected variables applied at a traffic signal and used for the comparative
analysis to compare results by Retina Net, Yolo V3, and Yolo Tiny on the same image