The problem
Plant diseases can be difficult to identify visually, especially where agricultural expertise is not immediately available.
KS
Building the experience
Loading projects, engineering decisions, and product details.
Back to projects
Artificial Intelligence Application
Final-year artificial-intelligence projectPlant Disease Detection
A machine-learning application that analyses uploaded crop images, predicts possible plant diseases, and presents accessible results through a Flask interface.
My role
Machine-learning and web developer
Core stack
Python · TensorFlow · Keras
plant-disease-detection.kwabenasoadwa.dev
The solution
I trained and integrated an image-classification model that accepts uploaded crop images, processes them, predicts a disease class, and displays the result through a Flask application.
Project overview
From practical need to complete product.
The Plant Disease Detection application connects a trained TensorFlow image-classification model to a web interface that allows users to upload crop images and receive predictions.
PythonTensorFlowKerasFlaskMachine Learning
Responsive product view
One interface, every screen.
The same product is presented across desktop, tablet, and phone frames to show how the interface adapts—not just how it looks in one screenshot.
desktop
tablet
phone
System architecture
A clear delivery path from the user interface through application logic to data and external services.
01Responsive image-upload interface
02Flask prediction API and preprocessing
03TensorFlow/Keras classification model
0418-class labelled image dataset
05Accessible prediction-result presentation
My responsibilities
- Dataset preparation and model training
- Image preprocessing and prediction integration
- Flask interface and upload workflow
- Responsive design, testing, and result presentation
Key outcomes
- Connected a trained model to a usable web workflow
- Supported eighteen plant-disease classes
- Made image selection and prediction usable across screen sizes
Key features
- Image-classification workflow
- TensorFlow model integration
- Flask-based prediction interface
- 18-class disease dataset
- Uploaded-image preview and results
Engineering challenges
- Preparing and organizing image datasets
- Connecting the trained model to Flask
- Processing uploaded images consistently
- Presenting model predictions clearly
Future improvements
- Add prediction confidence scores
- Add disease-treatment recommendations
- Improve mobile image capture
- Expand the supported crop classes
Lessons learned
What this project improved in my engineering process.
- 01Training and inference preprocessing must remain identical
- 02Prediction confidence and guidance improve user trust
- 03AI interfaces need clear limitations and accessible results
Next case study