Scribd
Upload
7 views9 pages

TPP Script New (Improved)

The document presents Drishti 2.0, an AI-powered real-time surveillance system aimed at enhancing women's safety by preventing incidents before they escalate. It combines object detection, action recognition, and instant alerting to ensure immediate intervention in public spaces. The system addresses the limitations of traditional surveillance, which are reactive and focus on post-incident analysis.

Uploaded by

aaryankuchekar06
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views9 pages

TPP Script New (Improved)

The document presents Drishti 2.0, an AI-powered real-time surveillance system aimed at enhancing women's safety by preventing incidents before they escalate. It combines object detection, action recognition, and instant alerting to ensure immediate intervention in public spaces. The system addresses the limitations of traditional surveillance, which are reactive and focus on post-incident analysis.

Uploaded by

aaryankuchekar06
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

AARYAN:

Opening Hook

Imagine walking home late at night. You’re alone, with only the
streetlights and CCTV cameras for company. But deep down, you
know that if something happens, those cameras will only record what
unfolds—they won’t stop it. How safe would you feel?

(Pause briefly)

This is the reality many women face daily. We’re here to present a
solution that changes this dynamic—a real-time surveillance system
designed not to just record but to prevent incidents before they
escalate.

Slide 1: Introduction

Good morning, esteemed judges and participants. I’m [Your Name],


and this is my colleague Aaryan Kuchekar. Today, we present our
technical paper, ‘AI-Powered Real-Time Surveillance for Enhancing
Women’s Safety.’
Our solution addresses a critical gap in traditional surveillance
systems. While these systems record footage, they lack real-time
intervention capabilities. This delay can mean the difference between
safety and danger.

Slide 2: Problem Statement

In today’s world, women’s safety is a critical concern, especially in


public spaces. While traditional surveillance systems have evolved,
they remain reactive, focusing on post-incident analysis.
For example, in incidents like the R.G. Kar Hospital assault and the
Badlapur School tragedy, cameras captured the events, but
intervention was delayed. Real-time detection could have prevented
the escalation of violence, highlighting the need for systems that act
before it’s too late.

Our paper highlights the need for real-time detection systems that
actively prevent escalation. Our proposed solution does exactly that—
detects threats in real-time and alerts authorities immediately.
YESH:
Slide 3: Why Real-Time Detection Matters

The key problem with traditional surveillance systems is their post-


event analysis. By the time the authorities are notified, it’s often too
late to stop the situation from escalating.

Our proposed system utilizes a combination of AI-driven


technologies, such as YOLO (You Only Look Once) for object
detection and ResNet-34 paired with LSTM (Long Short-Term
Memory) for action recognition. Together, these technologies identify
potential threats and send instant alerts to authorities. This proactive
approach is crucial for protecting women and vulnerable individuals
in public spaces.

Slide 4: Survey of Existing Systems

"Through an extensive survey of various research papers, we


gathered insights into existing surveillance systems and their
limitations."
"For instance, the first paper uses MobileNetV2 and a Telegram Bot
to send real-time alerts. It achieves 95% accuracy, but MobileNetV2
slows down detection, which is critical during emergencies."

"In the second paper, YOLOv5 with OpenCV achieves 99% accuracy
but lacks real-time notification capabilities, making it less useful in
urgent situations."

"The third paper combines YOLOv3 and Faster R-CNN, achieving


95% accuracy, but struggles in certain detection scenarios where
precision is vital."

The fourth paper uses YOLO with ResNet-34, achieving 82%


accuracy, but requires fine-tuning for more complex behavioral
detection, impacting its usability in dynamic environments.

"The fifth paper uses CNN and Deep Belief Networks, reaching 90%
accuracy, but lacks real-time capabilities altogether, which is essential
for public safety."

"Our goal is to combine the high-speed detection of YOLO with the


real-time alerting system of the first paper, offering comprehensive
public safety solutions focused on women’s safety."
AARYAN:
Slide 5: Our Proposed Solution – Drishti 2.0

Our proposed system, Drishti 2.0, integrates


1. YOLO for rapid object detection of people or dangerous items in
public spaces.
2. ResNet-34 + LSTM to recognize and analyze human actions,
distinguishing between normal behavior and potential threats, such as
harassment or violence.
3. Telegram Bot for real-time alerts that include images, timestamps,
and location data. This ensures that authorities are notified
immediately and can respond swiftly.

This seamless combination allows our system to detect threats like


harassment or violent behavior and notify the right people
immediately, ensuring faster response times and more effective
intervention.

Slide 6: System Workflow


Here’s how our system works in practice:

1. Step 1: YOLO detects objects like people or weapons in live


surveillance feeds.
2. Step 2: ResNet-34 and LSTM recognize specific actions—such as
an aggressive approach or harassment.
3. Step 3: Upon detecting a threat, the Telegram Bot sends an
immediate alert with visual data, timestamps, and location to the
authorities, enabling a rapid response.

This process happens in real-time, ensuring immediate intervention to


prevent incidents from escalating.

YESH:
Slide 7: Key Advantages of Our System

Here’s what sets Drishti 2.0 apart:


3. Real-Time Alerts: Our integration with Telegram Bots ensures
faster, real-time alerts, reducing response times compared to email-
based notifications.
2. Women’s Safety Focus: Drishti 2.0 is one of the first surveillance
systems specifically designed for preventing violence against women
in public spaces.

1. Proactive Detection: Unlike traditional systems, we prevent


incidents before they escalate by detecting threats in real-time.
(“(saleable read from ppt)
faster response
Contextual”)
4. Cost-Efficiency: The system integrates with existing CCTV
infrastructure, minimizing hardware upgrade costs.
5. Edge Computing: With low latency and high-speed processing,
Drishti 2.0 can be deployed at scale without overloading
infrastructure.

Slide 8: Challenges and Future Directions

While our system addresses many key issues, there are challenges to
consider:
1. High Computational Demand: Technologies like YOLO, ResNet-
34, and LSTM require substantial resources. We aim to optimize
these models for edge computing to reduce hardware demands.
2. Dataset Limitations: Currently, available datasets for women’s
safety are limited in size and diversity. This restricts the detection
accuracy of more subtle forms of harassment. Moving forward, we
aim to build larger, specialized datasets to improve detection accuracy
and system robustness.

AARYAN:
Slide 9: Conclusion

In conclusion, Drishti 2.0 offers a groundbreaking solution for real-


time surveillance, focusing on preventing incidents before they
happen. By integrating object detection, action recognition, and
instant alerting systems, we ensure faster responses and safer public
spaces.

Our system is scalable, efficient, and adaptable for various


environments, making it a vital tool in enhancing public safety,
particularly for women.
Closing the Presentation

Together, we have the power to transform public safety. Our system


doesn’t just record—it acts. Drishti 2.0 promises a future where
surveillance is not passive but proactive, ensuring that women and
other vulnerable individuals are protected in real time.

Thank you for your attention. We’re happy to take any questions you
may have.

You might also like