Meet The Developer

Prepare to meet our extraordinary team. We are a diverse group of

talented developers and forward-thinking innovators, united by a
common goal: to redefine safety and emergency response through
advanced computer vision technology. With a passion for creating
positive change, we are dedicated to enhancing accident detection,
assessing injury severity, and ensuring swift medical attention. Join us on
this transformative journey towards a safer future, where our vision sets
new standards in security and well-being.

Team name: Antardhwani

Team members:
-Gagan Sharma
-Ishant Verma
-Ronit Ranjan Tripathy
-Varun Gupta
Problem Statement
● Accidents are an unfortunate reality in our society, encompassing a wide spectrum from common traffic collisions to
workplace mishaps. However, within this challenging landscape lies a deeply concerning issue:

● Critical Delays in Medical Assistance: It is all too common for individuals involved in accidents to endure significant
and sometimes tragic delays in receiving the immediate medical attention they desperately need. These delays can
lead to avoidable fatalities, casting a pall of grief over victims and their families, who are left grappling with the
devastating realization that timely intervention might have saved lives.

● Subjective Injury Assessment: Compounding the problem is the subjective nature of injury assessment. Conventional
methods for gauging injury severity can be both subjective and time-consuming, adding further complexity to the
situation and exacerbating the consequences.

● Loss of Lives Due to Delay: The most heart-wrenching aspect of this situation is the loss of lives that could have been
preserved. Accident victims often find themselves in dire circumstances, and their fate is frequently attributed to the
distressing notion of 'lateness' or the challenges faced by emergency services in arriving promptly.

● This multifaceted issue underscores the urgent need to reevaluate and enhance the efficiency and accuracy of our
emergency response systems. Our unwavering commitment is to prevent avoidable loss of life and to alleviate the
suffering experienced by individuals and communities affected by these tragic events.
 ● Prepare to be amazed by a solution that breaks free from
conventions and provides a vital lifeline during critical moments. Our
visionary team has crafted an exceptional response to the pressing

Solution
challenges of accident management and emergency care—a solution
that is genuinely unique and poised to redefine industry norms.

Understanding
● Accident Detection with Real-time Severity Assessment: Our state-of-
the-art computer vision system is a game-changer in accident
detection. It boasts unmatched capabilities, swiftly and accurately
identifying accidents while providing real-time assessments of injury
severity. The standout feature of our approach is its precision and
speed, setting a new standard for accident response.

● SOS Generation and Immediate Alert: What truly sets our solution
apart is its ability to autonomously generate SOS alerts. When an
accident is detected, our system springs into action, instantly
triggering alerts to emergency services and designated contacts
which will be attached with the location of mishap with photos of it.
It's a level of responsiveness that's unparalleled, reinforcing our
commitment to ensuring lives are safeguarded promptly.Our SOS
system not only notifies medical agencies but also relays accident
details to the victim's family, expediting their arrival at the scene.

● Our cutting-edge computer vision and machine learning models

empower us to effectively classify accidents, determining whether
victims require immediate medical attention. Using a sophisticated
approach, our system evaluates various critical factors within each
event, enabling us to accurately identify and prioritize individuals in
dire need of urgent medical assistance.With our model we were able
to achieve high accuracy and also we are able to generate output
just in the moment accidents happen.
 Our Approach 01
App Logic
Real time accident detection in background

• Our computer vision system swiftly and accurately identifies accidents while assessing injury 02 Evaluate the severity of detected accidents based on
severity in real-time, setting new standards for precision and speed in accident response factors like impact force, vehicle damage, and the
number of involved vehicle
• Our solution autonomously generates SOS alerts upon accident detection, instantly notifying
emergency services and designated contacts with precise location information and 03 If an accident is detected and its sevarity meets
photographic evidence, ensuring rapid and informed responses
predefined criteria for urgent medical attention, initiate
SOS generation
• We leverage the capabilities of YOLO (You Only Look Once) and OpenCV for real-time
04 Generate a message for family members to provide
object detection, ensuring swift accident recognition. immediate assistance and peace of mind to the victim

• Our system operates in real-time, enabling immediate responses to accidents as they occur

• Our solution incorporates neural networks for advanced accident classification Technology stack:
• We integrate RTO APIs to gather victim information from their vehicle's number plate
YOLO,OpenCV,RTO api,Google Maps api,Convolutional Neural Networks
(CNNs),tensorflow,numpy,keras,PyTorch,Faster R-CNN,imgaug.
• We collaborate with pre-installed CCTV cameras on roadways in various areas to expand our
accident detection network
Key Features:
• We design and implement our own SOS system for generating alerts, ensuring optimal
response times Real time accident detection,SOS generation,accident
classification,high precision,instantaneous response,Severity
• By utilizing Google Maps APIs, we identify the five closest hospitals to the accident scene.
The first hospital to accept the SOS will receive the patient, prioritizing proximity for swift care Assessment,Family Notification,Scalability,Location Awareness

Assumptions:
The main assumption is that accident must be caught in the camera and
result will depend on the quality of video so the video quality must be more
than 480p
PPE In Action
Video:-

Link to the video -

PROGRESS WE MADE

In the beginning, our system could solely identify accidents and trigger SOS notifications. However,
following several significant enhancements to our model.
We are now able to alert hospital about these unfortunate events related to vehicle accidents also
we can now differentiate accidents into three distinct categories:
1. Normal
2. Moderate
3. Fatal Accident
The imperative for these improvements was exceptionally high, as our central mission is to offer
timely aid to those in critical need, safeguarding lives hanging in the balance. With these
enhancements in place, we can promptly generate SOS alerts for accidents classified as "fatal,"
guaranteeing swift medical assistance when it is most critical.
Thank You