Leaf Disease Detection Using
Deep Learning
Final Review Presentation
Department of Computer Science
and Engineering
Abstract
• This project uses Convolutional Neural
Networks (CNNs) to detect and classify leaf
diseases from different plant species. It
improves early detection, reduces crop loss,
and promotes sustainable agriculture. The
system includes a GUI using Tkinter and uses
TensorFlow, OpenCV, and other tools.
System Requirement
HARDWARE REQUIREMENT
Processor : Intel Core i3 or higher
RAM : Minimum 8 GB
Hard Disk : At least 100 GB
Key Board : Standard IOS Keys
Monitor : 14’ color monitor
SOFTWARE REQUIREMENT
Operating System : MS-Windows
Programming : Python 3.x
Framework : Keras
Problem Definition
• Traditional disease detection methods are
time-consuming and require expert
intervention. There is a need for an
automated, accurate, and efficient system for
real-time leaf disease detection.
Proposed System
• A CNN-based deep learning model is proposed
to classify leaf diseases. The system uses
image preprocessing, model prediction, and a
GUI to enable user-friendly disease diagnosis.
ADVANTAGE AND DISADVANTAGE
• Leaf disease detection has advantages like
early detection, increased crop yield, and
reduced chemical usage, promoting healthier
plants and environmental sustainability.
However, it also has disadvantages, including
limited accuracy, dependence on high-quality
images, and computational requirements,
which can affect its
effectiveness and efficiency.
Software Tools Used
• • Python 3.x
• • TensorFlow & Keras
• • OpenCV
• • Tkinter
• • NumPy & Pandas
• • Matplotlib
• • JSON & H5 File Handling
System Flow Diagram
• Includes Level 0, Level 1, and Level 2 diagrams
showing data flow and system interactions.
Input and Output Design
• Input: Leaf images from different plant
species.
• Output: Predicted disease name with
confidence score displayed through GUI.
Testing and Implementation
• The system underwent unit testing,
integration testing, and validation with sample
datasets. It was successfully implemented
with a user-friendly GUI.
OUTPUT
Conclusion
• The system provides a scalable, accurate, and
accessible solution for leaf disease detection,
empowering farmers and researchers to take
proactive measures in plant health
management.
Future Enhancement
• • Expand the dataset with more plant species
• • Improve accuracy using advanced deep
learning models
• • Deploy as a mobile application for real-time
diagnosis in the field
