Fracture detection in

pediatric wrist trauma

X‑ray images using
YOLOv8 algorithm
Published on November 16, 2023, in Scientific
Reports journal, which is classified as a Q1
journal.
Scientific Reports journal
Scientific reports: has a 2-year impact factor of 3.8
(2023), and is the 5th most-cited journal in the world, with
more than 734,000 citations in 2023
1. Journal Impact Factor (JIF):
• 2023 Value: 3.8
The JIF represents the average number of citations received per paper published
in the journal during the previous two years. For example, if 200 articles
published in 2021 and 2022 were cited 760 times in 2023, the JIF would be
760 / 200 = 3.8. It’s a common measure of a journal’s influence.
Scientific Reports journal
2. 5-Year Journal Impact Factor:
• Value: 4.3
This metric is like the JIF but considers citations over the past five years instead
of two, providing a more extended perspective on the journal's impact.

3. Immediacy Index:
• Value: 0.8
This measures how quickly, on average, articles in the journal are cited within the
year of their publication. A higher index indicates faster recognition and use of the
research.
Scientific Reports journal
4. Eigenfactor® Score:
• Value: 0.90790
This metric evaluates the journal's overall importance within the scientific
community, considering the number of incoming citations and the quality of the
citing journals. Unlike the JIF, it doesn’t count self-citations and gives more
weight to influential journals.

5. Article Influence Score:

•Value: 1.1
This score measures the average influence of each article in the journal over five
years. A score greater than 1.0 indicates that the journal’s articles are more
influential than average in their field.
 Table of contents

01 02 03
Introduction Problem Use of Deep
learning in Bone
Fracture Detection

04 05 06
Name of the Name of the Name of the
section 4 section 5 section 6
01
Introductio
n
In hospital emergency rooms, radiologists
are often asked to examine patients with
fractures in various parts of the body,
such as the wrist and arm.
Fractures can generally be classified as open or
closed:
● open fractures occurring when the bone
pierces the skin
● closed fractures occurring when the skin
remains intact despite the broken bone

Before performing surgery, the surgeon must

inquire about the medical history of the patients
and conduct a thorough examination to diagnose
fracture
02

Problem
In prestigious hospitals of developed areas,
there are many experienced radiologists who
are capable of correctly analyzing X-ray images.

The problem: is that in some small hospitals of

underdeveloped regions, there are only young
and inexperienced surgeons who may be unable
to correctly interpret X-ray images. Therefore, a
shortage of radiologists would seriously
jeopardize timely patient care.

Specifically, some hospitals in Africa have even

limited access to specialist reports, which badly
affects the probability of the success of surgery.
According to the survey, the percentage of X-ray
images misinterpreted have reached 26%.
 recent medical
imaging

Computed Magnetic
Tomography Resonance Imaging X-ray
(CT), (MRI)
is the most
offering faster imaging ideal for widely used
than MRI and better detailed device due to
resolution than X-rays, but assessments of
its cost-
they use ionizing radiation, organs, effectiveness
which poses a higher risk
 03
Introduce Deep
Learning in Bone
Fracture
Detection
The solution of the previous problem is to use the deep
learning models to automate Bone Fracture Detection Process.
With the advancement of deep learning, neural network
models have been introduced in medical image processing.
In recent years, researchers have started to apply object
detection models to fracture detection, which is a popular
research topic in computer vision (CV).

Deep learning methods in the field of object detection are

divided into:
1. one-stage algorithms
2. two-stage algorithms
In this paper They first use YOLOv8 algorithm which is one-
stage algorithms
 04
Yolo Architecture
Yolo is stands for you only look once.

With the advancement of deep learning, neural network

models have been introduced in medical image processing.
In recent years, researchers have started to apply object
detection models to fracture detection, which is a popular
research topic in computer vision (CV).

Deep learning methods in the field of object detection are

divided into:
1. one-stage algorithms
2. two-stage algorithms
In this paper They first use YOLOv8 algorithm which is one-
stage algorithms
05

Dataset
 GRAZPEDWRI-DX Dataset
Provided by Medical University of Graz, consists of 20,327 X-ray
images of wrist trauma in children. These images were collected
from 6,091 patients between 2008 and 2018 by multiple pediatric
radiologists at the Department of Pediatric Surgery of the
University Hospital Graz. The images are annotated in 9 different
classes by placing bounding boxes on them.
06
Experimen
ts
 Data Splitting
They randomly divide the 20,327 X-ray images of the
GRAZPEDWRI-DX dataset into:
1. Training set 70% (14,204 Image)
2. Validation set 20% (4,094 Image)
3. Test set 10% (2,029 Image)
 Data Augmentation
They used Data Augmentation to extend dataset and increase the
number of images in training set from 14,204 X-ray images to
28,408 X-ray images.

They Generate new images using equation (1), Where Input1 and
Input2 are the two input images of the same size respectively, α
represents the weight assigned to the first input image, β denotes
the weight assigned to the second input image, and γ represents
the scalar value added to each sum.
(1)
 Data Augmentation
Their purpose is to adjust the contrast and brightness of the original input image; They take the
same image as Input1 and Input2 respectively and set β to 0. The value of α and γ represent the
proportion of the contrast and the brightness of the image, respectively.

After comparing different settings, They finally decided to set α to 1.2 and γ to 30 to avoid the
output image being too bright.
 Six ideas
Mars Venus Neptune
Despite being red, Venus has a beautiful Neptune is the
Mars is actually a cold name and is the farthest planet from
place. It's full of iron second planet from the Sun. It's also the
oxide dust, which the Sun. It’s terribly fourth-largest planet
gives the planet its hot, even hotter than by diameter in the
reddish cast Mercury Solar System

Earth Saturn Jupiter

Earth is the third Saturn is a gas giant Jupiter is a gas giant
planet from the Sun and has rings. It's and the biggest
and the only one that composed mostly of planet in the Solar
harbors life in the hydrogen and helium. System. It's the
Solar System. We all It’s the sixth planet fourth-brightest
live on this planet from the Sun object in the night sky
 50,000
This is the number of total visitors
we had last month

20,000 5,000
This is the number of patients we This is the number of transplants
had last month performed so far
A picture is worth a thousand words
 95%
Mercury is the closest planet
to the Sun and the smallest
one in the Solar System—it’s
only a bit larger than the
Moon
 Our statistics

Understanding
the numbers
Healthcare center statistics
refer to the numerical data
and metrics that provide
insight into the
performance, outcomes,
and patient demographics
of a healthcare facility.
These statistics are used to
inform decision-making and
identify areas for
improvement in the
delivery of healthcare
Follow the link in the graph to modify its data and then paste the new one here. For more info, click here
services
 Areas we cover

Jupiter Mercury Venus Mars

Area 1 ✔ ✖ ✔ ✖

Area 2 ✖ ✔ ✔ ✖

Area 3 ✔ ✖ ✔ ✔

Area 4 ✖ ✔ ✖ ✔
 Milestones reached

01 Venus is very hot

05 Mars is very cold

02 Saturn is a gas giant

06 The Moon is a
satellite

03 Mercury is very small

07 Earth has life

04 Jupiter is big
08 Neptune is an ice
giant
Photo showcase
You can replace the images on the screen with
your own work. Just right-click on any of them and
select “Replace image”
 Data and statistics
Healthcare Satisfactio Patients Number of patients seen each
service n rate (%) surveyed
month
Service 1 92 1,000
Service 2 85 500
Service 3 90 750

Xxx billion
Despite being red, Mars is
actually a cold place
Xh xmin
Venus is extremely hot, even
more than Mercury
Xx
Follow the link in the graph to modify its data and
Earth is the planet on which we then paste the new one here.
all live For more info, click here
 Testimonials

“Mercury is the closest

“Neptune is the fourth- planet to the Sun and the “Saturn is the ringed planet.
largest planet by diameter in smallest one in the Solar It’s composed mostly of
the entire Solar System” System” hydrogen and helium”

Jenna Doe, 25 John James, 56 Ana Adams, 39