Skip to content

42 AI Innovative computer vision project utilizing leaf image analysis for disease recognition

Notifications You must be signed in to change notification settings

qurobert/Leaffliction

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

49 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Leaffliction 🥬🩺

Image classification & disease recognition on plant leaves

A complete computer‑vision pipeline that balances your dataset, augments images, extracts key features (PlantCV), trains a CNN with TensorFlow / Keras, signs the resulting artefacts, and predicts diseases from new leaf photos.


✨ Features

Stage Script What it does
Dataset insight src/Distribution.py Draw pie / bar charts to visualise class imbalance
Augmentation src/Augmentation.py Flip, rotate, crop, blur, contrast & projective transforms
Pre‑processing src/train_preprocessing.py Balances each class, calls augmentation, removes background
Image transforms src/Transformation.py Masking, landmarks, histograms & other PlantCV goodies
Training src/train.py Splits data, trains a CNN, logs metrics, zips model + data + SHA‑256 signature
Inference src/predict.py Predict a single image or a whole directory and display results

🚀 Quick start

# 1 – create and activate a virtual environment
python -m venv .venv && source .venv/bin/activate

# 2 – install dependencies
pip install -r requirements.txt   # opencv-python, tensorflow, plantcv, matplotlib, tqdm …

# 3 – inspect your raw dataset
python src/Distribution.py data/raw

# 4 – balance & augment (optional standalone run)
python src/train_preprocessing.py data/raw --augmented_dir data/augmented

# 5 – train the model
python src/train.py data/raw \
     --augmented_dir data/augmented \
     --mask_directory data/augmented_mask \
     --model_dir data/model

# 6 – predict
python src/predict.py path/to/leaf.jpg --model_dir data/model
# or an entire folder
python src/predict.py data/test_images --model_dir data/model

📈 Results

With the default 80 / 20 split the network reaches > 90 % validation accuracy (10 epochs, 256×256 inputs). Loss/accuracy curves are displayed automatically at the end of training.

About

42 AI Innovative computer vision project utilizing leaf image analysis for disease recognition

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages