YOLO-BASED OBJECT
DETECTION
Presented By:
[Link]
160122771001
B.E. AI&DS-(I1)
INTRODUCTION
Object detection identifies objects and their locations in images and
videos.
Real-time detection requires high speed and low latency.
Traditional object detection methods are slow and computationally
expensive.
YOLO enables fast, real-time detection using a single-stage
architecture.
YOLO provides an effective balance between detection speed and
accuracy.
LITERATURE SURVEY
YOLO Version /
Year Authors Paper Title Problem Addressed Key Contribution
Algorithm Used
Real-time Object Detection Provided a comprehensive
and Segmentation Need for fast real-time object analysis of YOLO evolution,
2023 Kang & Kim YOLOv1 – YOLOv8 (Review)
Technology: An Analysis of detection and segmentation performance, strengths, and
the YOLO Algorithm limitations
Improved small-object
Improved YOLOv7 Algorithm
Modified YOLOv7 (YOLOv7- Small object detection in UAV detection accuracy with
2024 Xinmin Li et al. for Small Object Detection in
UAV) images reduced computational cost
UAV Image Scenarios
for real-time UAV deployment
Explained YOLO’s real-time
Enhancing Real-time Object
Efficient real-time object capability and compared it
2023 Lavanya & Pande Detection with YOLO YOLOv2, YOLOv3, YOLOv4
detection with traditional detection
Algorithm
methods
Demonstrated YOLO’s
YOLO in Suspicious Human YOLOv3 – YOLOv7 Real-time surveillance and effectiveness in intelligent
2026 Widodo et al.
Activity Recognition (Integrated systems) suspicious activity detection surveillance and security
systems
ARCHITECTURE
YOLOV1
YOLOv1 is the first version of the YOLO algorithm,
introduced by Joseph Redmon et al. in 2016. It
was the first object detection model capable of
real-time detection using a single neural network.
YOLOV8
YOLOv8 is the latest YOLO generation designed
for real-time object detection with higher
accuracy, simpler design, and faster inference. It
improves on earlier YOLO versions by simplifying
the detection head and optimizing training and
deployment.
YOLOV7
YOLOV7 IS A REAL-TIME OBJECT
DETECTION MODEL DESIGNED TO
DETECT OBJECTS VERY FAST AND
VERY ACCURATELY. IT IS ONE OF
THE MOST EFFICIENT YOLO
VERSIONS AND IS WIDELY USED IN
REAL-WORLD APPLICATIONS LIKE
SURVEILLANCE, UAVS, AND
AUTONOMOUS SYSTEMS.
CONSOLIDATION OF AVAILABLE
INFORMATION
YOLO analyzes the entire image for better contextual
understanding.
Eliminates complex region proposal steps used in traditional
models.
Optimized for deployment on both high-end and low-resource
devices.
Uses advanced feature fusion to improve detection quality.
Designed to maintain efficiency with reduced memory and
computation cost.
Strengths
High detection speed
End-to-end learning approach
Low computational cost
Suitable for real-time systems (UAVs, surveillance)
Challenges
Difficulty detecting small and overlapping objects
Performance issues in crowded or complex scenes
Requires large, well-labeled datasets
YOLO ARCHITECTURE
The image shows YOLO detecting
multiple objects (dog, bicycle,
truck) in a single image using
bounding boxes and class labels.
All objects are detected
simultaneously in one forward
pass, demonstrating YOLO’s real-
time object detection capability.
SUMMARY
YOLO is a powerful deep learning–based algorithm for object
detection.
It enables real-time detection by performing object localization and
classification in a single pass.
YOLO achieves a good balance between detection speed and accuracy.
It is widely applied in real-time systems such as UAVs, surveillance, and
intelligent monitoring.
Recent research focuses on improving YOLO performance in
challenging scenarios like small object detection and crowded
environments.
CONCLUSION
YOLO-based object detection is highly suitable for real-time
applications.
Its single-stage architecture makes it faster than traditional
detection methods.
Continuous improvements in YOLO models enhance accuracy
without compromising speed.
YOLO remains one of the leading frameworks for real-time object
detection.
Future advancements are expected to further expand YOLO’s
applications in intelligent systems.
REFERENCES
1. Xinmin Li et al., Improved YOLOv7 Algorithm for Small Object Detection
in UAV Image Scenarios, Applied Sciences, 2024.
2. Gudala Lavanya, S. D. Pande, Enhancing Real-time Object Detection with
YOLO Algorithm, EAI Endorsed Transactions on Internet of Things, 2023.
3. Yohanes Bowo Widodo et al., YOLO in Suspicious Human Activity
Recognition for Intelligent Environmental Security Systems, Journal of
Theoretical and Applied Information Technology, 2026.
4. Chang Ho Kang, Sun Young Kim, Real-time Object Detection and
Segmentation Technology: An Analysis of the YOLO Algorithm, JMST
Advances (Springer), 2023.
Thank you